avance_notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "17404701",
   "metadata": {},
   "source": [
    "# Forecasting recovery rates for debt collection agency: Avance BPO\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d3dafb41",
   "metadata": {},
   "source": [
    "\n",
    "A non-performing loan (NPL) is a loan that is in default due to the fact that the borrower has not made the scheduled payments for a specified period.\n",
    "\n",
    "Non-performing loans are purchased from the bank by a debt collection agency which must prioritize which users to prompt first. One of the most important factors regarding this decision is the recovery rate: the percentage of exposure that can be recovered from each borrower through the debt collection process. The recovery rates achievable by the debt collector are unknown at the time of purchase of the portfolio and need to be predicted. \n",
    "\n",
    "Based on the information given by the bank, the objective of this model is to predict the recovery rates on non-performing loans  (NLP) using a private database from a Colombian debt collection agency called Avance BPO."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd025487",
   "metadata": {},
   "source": [
    "## Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "32e5a3d1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/proctor/.pyenv/versions/3.8.12/envs/lewagon/lib/python3.8/site-packages/scipy/__init__.py:146: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.23.3\n",
      "  warnings.warn(f\"A NumPy version >={np_minversion} and <{np_maxversion}\"\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cd5634e3",
   "metadata": {},
   "source": [
    "## Importing Datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6a18ddef",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>FechaCaptura</th>\n",
       "      <th>Credito</th>\n",
       "      <th>FechaAPagar</th>\n",
       "      <th>VRCancelar</th>\n",
       "      <th>PuntoPago</th>\n",
       "      <th>TipoPago</th>\n",
       "      <th>Asesor</th>\n",
       "      <th>VrTotalObligacion</th>\n",
       "      <th>VrTotalNegociado</th>\n",
       "      <th>Condonacion</th>\n",
       "      <th>TipoPago2</th>\n",
       "      <th>Cuotas</th>\n",
       "      <th>VrCuota1</th>\n",
       "      <th>VrOtrasCuotas</th>\n",
       "      <th>CuotaNumero</th>\n",
       "      <th>Estado</th>\n",
       "      <th>FechaPago</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-11-08</td>\n",
       "      <td>904141001352</td>\n",
       "      <td>2016-11-30</td>\n",
       "      <td>100000</td>\n",
       "      <td>CDS o Efecty</td>\n",
       "      <td>Pago Parcial</td>\n",
       "      <td>Si</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1900-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-11-23</td>\n",
       "      <td>904141001295</td>\n",
       "      <td>2016-12-21</td>\n",
       "      <td>260000</td>\n",
       "      <td>Oficina Fundaci¢n de la Mujer</td>\n",
       "      <td>Pago Parcial con Condonaci¢n</td>\n",
       "      <td>Si</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1900-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-11-28</td>\n",
       "      <td>904141000816</td>\n",
       "      <td>2016-12-15</td>\n",
       "      <td>844000</td>\n",
       "      <td>Oficina Fundaci¢n de la Mujer</td>\n",
       "      <td>Pago Parcial con Condonaci¢n</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1900-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-11-19</td>\n",
       "      <td>904141000465</td>\n",
       "      <td>2016-11-30</td>\n",
       "      <td>75000</td>\n",
       "      <td>CDS o Efecty</td>\n",
       "      <td>Pago Parcial con Condonaci¢n</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1900-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-11-09</td>\n",
       "      <td>904141000343</td>\n",
       "      <td>2016-11-30</td>\n",
       "      <td>55000</td>\n",
       "      <td>CDS o Efecty</td>\n",
       "      <td>Pago Parcial con Condonaci¢n</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1900-01-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  FechaCaptura       Credito FechaAPagar  VRCancelar  \\\n",
       "0   2016-11-08  904141001352  2016-11-30      100000   \n",
       "1   2016-11-23  904141001295  2016-12-21      260000   \n",
       "2   2016-11-28  904141000816  2016-12-15      844000   \n",
       "3   2016-11-19  904141000465  2016-11-30       75000   \n",
       "4   2016-11-09  904141000343  2016-11-30       55000   \n",
       "\n",
       "                       PuntoPago                      TipoPago Asesor  \\\n",
       "0                   CDS o Efecty                  Pago Parcial     Si   \n",
       "1  Oficina Fundaci¢n de la Mujer  Pago Parcial con Condonaci¢n     Si   \n",
       "2  Oficina Fundaci¢n de la Mujer  Pago Parcial con Condonaci¢n    NaN   \n",
       "3                   CDS o Efecty  Pago Parcial con Condonaci¢n    NaN   \n",
       "4                   CDS o Efecty  Pago Parcial con Condonaci¢n    NaN   \n",
       "\n",
       "   VrTotalObligacion  VrTotalNegociado  Condonacion  TipoPago2  Cuotas  \\\n",
       "0                  0                 0            0        NaN       0   \n",
       "1                  0                 0            0        NaN       0   \n",
       "2                  0                 0            0        NaN       0   \n",
       "3                  0                 0            0        NaN       0   \n",
       "4                  0                 0            0        NaN       0   \n",
       "\n",
       "   VrCuota1  VrOtrasCuotas  CuotaNumero Estado  FechaPago  \n",
       "0         0              0            0    NaN 1900-01-01  \n",
       "1         0              0            0    NaN 1900-01-01  \n",
       "2         0              0            0    NaN 1900-01-01  \n",
       "3         0              0            0    NaN 1900-01-01  \n",
       "4         0              0            0    NaN 1900-01-01  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# df = pd.read_excel('datos_avance.xlxs', sheet_name=None)\n",
    "df = pd.read_excel('datos_avance.xlsx') \n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "5db7c7d1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['acuerdos', 'payment_params', 'pagos', 'asignacion_fdem', 'resumen_creditos', 'gestiones_2022']\n"
     ]
    }
   ],
   "source": [
    "dfname=pd.ExcelFile('datos_avance.xlsx')\n",
    "print(dfname.sheet_names)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "890183d4",
   "metadata": {},
   "source": [
    "### Pagos Dataset: For extracting Target Value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "27e25a13",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MesCargue</th>\n",
       "      <th>Codigo</th>\n",
       "      <th>Agencia</th>\n",
       "      <th>Credito</th>\n",
       "      <th>Estado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "      <th>DiasMoraFindeMes</th>\n",
       "      <th>SaldoActual</th>\n",
       "      <th>DiasmoraActual</th>\n",
       "      <th>FechaValor</th>\n",
       "      <th>...</th>\n",
       "      <th>MORA</th>\n",
       "      <th>PAGOSOTROS</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>VRgoech</th>\n",
       "      <th>CONDCAPITAL</th>\n",
       "      <th>CONDINTERES</th>\n",
       "      <th>CONDMORA</th>\n",
       "      <th>DiasMoraPago</th>\n",
       "      <th>Reverso</th>\n",
       "      <th>Concepto</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>101</td>\n",
       "      <td>BUCARAMANGA AGENCIA PRINCIPAL</td>\n",
       "      <td>101101002728</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>321631</td>\n",
       "      <td>2032</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-18</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>10556</td>\n",
       "      <td>332187</td>\n",
       "      <td>321631</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>594634</td>\n",
       "      <td>2050</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>101</td>\n",
       "      <td>BUCARAMANGA AGENCIA PRINCIPAL</td>\n",
       "      <td>101104101267</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>149159</td>\n",
       "      <td>2142</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-31</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>4182</td>\n",
       "      <td>153341</td>\n",
       "      <td>149159</td>\n",
       "      <td>0</td>\n",
       "      <td>3498</td>\n",
       "      <td>356515</td>\n",
       "      <td>2172</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>101</td>\n",
       "      <td>BUCARAMANGA AGENCIA PRINCIPAL</td>\n",
       "      <td>101131041847</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>384108</td>\n",
       "      <td>490</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-20</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>9799</td>\n",
       "      <td>393907</td>\n",
       "      <td>384108</td>\n",
       "      <td>0</td>\n",
       "      <td>20307</td>\n",
       "      <td>233846</td>\n",
       "      <td>510</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>102</td>\n",
       "      <td>BUCARAMANGA AGENCIA ASTURIAS</td>\n",
       "      <td>102141028221</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>291780</td>\n",
       "      <td>706</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-21</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>20420</td>\n",
       "      <td>312200</td>\n",
       "      <td>291780</td>\n",
       "      <td>0</td>\n",
       "      <td>30678</td>\n",
       "      <td>226165</td>\n",
       "      <td>727</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>102</td>\n",
       "      <td>BUCARAMANGA AGENCIA ASTURIAS</td>\n",
       "      <td>102141028298</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>1363613</td>\n",
       "      <td>803</td>\n",
       "      <td>1322611</td>\n",
       "      <td>803</td>\n",
       "      <td>2016-10-18</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>12549</td>\n",
       "      <td>100000</td>\n",
       "      <td>87451</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>52487</td>\n",
       "      <td>821</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MesCargue  Codigo                        Agencia       Credito     Estado  \\\n",
       "0   2016-10     101  BUCARAMANGA AGENCIA PRINCIPAL  101101002728  Castigado   \n",
       "1   2016-10     101  BUCARAMANGA AGENCIA PRINCIPAL  101104101267  Castigado   \n",
       "2   2016-10     101  BUCARAMANGA AGENCIA PRINCIPAL  101131041847  Castigado   \n",
       "3   2016-10     102   BUCARAMANGA AGENCIA ASTURIAS  102141028221  Castigado   \n",
       "4   2016-10     102   BUCARAMANGA AGENCIA ASTURIAS  102141028298  Castigado   \n",
       "\n",
       "   SaldoFindeMes  DiasMoraFindeMes  SaldoActual  DiasmoraActual FechaValor  \\\n",
       "0         321631              2032            0               0 2016-10-18   \n",
       "1         149159              2142            0               0 2016-10-31   \n",
       "2         384108               490            0               0 2016-10-20   \n",
       "3         291780               706            0               0 2016-10-21   \n",
       "4        1363613               803      1322611             803 2016-10-18   \n",
       "\n",
       "   ... MORA PAGOSOTROS  TotalCancelado  VRgoech CONDCAPITAL  CONDINTERES  \\\n",
       "0  ...    0      10556          332187   321631           0            0   \n",
       "1  ...    0       4182          153341   149159           0         3498   \n",
       "2  ...    0       9799          393907   384108           0        20307   \n",
       "3  ...    0      20420          312200   291780           0        30678   \n",
       "4  ...    0      12549          100000    87451           0            0   \n",
       "\n",
       "   CONDMORA  DiasMoraPago  Reverso  Concepto  \n",
       "0    594634          2050      NaN       NaN  \n",
       "1    356515          2172      NaN       NaN  \n",
       "2    233846           510      NaN       NaN  \n",
       "3    226165           727      NaN       NaN  \n",
       "4     52487           821      NaN       NaN  \n",
       "\n",
       "[5 rows x 27 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pagos =pd.read_excel(f'datos_avance.xlsx',sheet_name='pagos')\n",
    "df_pagos.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "91a68939",
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'df_out2' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m/tmp/ipykernel_577/3555514388.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mdf_out2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdf_out2\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'Credito'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m\u001b[0;36m108171029529\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'df_out2' is not defined"
     ]
    }
   ],
   "source": [
    "df_out2.loc[df_out2['Credito'] ==108171029529]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 602,
   "id": "c7e334a9",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# df_acuerdos =pd.read_excel(f'datos_avance.xlsx',sheet_name='acuerdos')\n",
    "# df_acuerdos.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6a9bdc69",
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_acuerdos.to_csv(\"acuerdos.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 603,
   "id": "28dab7f6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# df_payment_params =pd.read_excel(f'datos_avance.xlsx',sheet_name='payment_params')\n",
    "# df_payment_params.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "7c258d51",
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_payment_params.to_csv(\"payment_params.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "dea8a2d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_pagos.to_csv(\"pagos.csv\",index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "943f0a11",
   "metadata": {},
   "source": [
    "### Bank NPL Dataset "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "80ebab69",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>obligacion</th>\n",
       "      <th>Titular</th>\n",
       "      <th>agencia</th>\n",
       "      <th>tipo</th>\n",
       "      <th>primera_asignacion</th>\n",
       "      <th>fecha_asignacion</th>\n",
       "      <th>fecha_desasignacion</th>\n",
       "      <th>nombre</th>\n",
       "      <th>ciudad</th>\n",
       "      <th>tel_casa</th>\n",
       "      <th>...</th>\n",
       "      <th>inter_corri</th>\n",
       "      <th>inter_mora</th>\n",
       "      <th>hyc</th>\n",
       "      <th>seguro</th>\n",
       "      <th>iva</th>\n",
       "      <th>valor_cuota</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>201917264</td>\n",
       "      <td>1</td>\n",
       "      <td>201</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2019-07-22 00:00:00</td>\n",
       "      <td>SALAZAR ECHAVARRIA   CARMEN ADRIANA</td>\n",
       "      <td>CUCUTA</td>\n",
       "      <td>75702725</td>\n",
       "      <td>...</td>\n",
       "      <td>29994</td>\n",
       "      <td>2397521</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10191142053</td>\n",
       "      <td>2</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2019-07-22 00:00:00</td>\n",
       "      <td>GONZALEZ GONZALEZ   EFRAIN</td>\n",
       "      <td>BUCARAMANGA</td>\n",
       "      <td>76731342</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>118354</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>11</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10191144858</td>\n",
       "      <td>3</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2019-07-22 00:00:00</td>\n",
       "      <td>LANDAZABAL SANTANDER   MILENA</td>\n",
       "      <td>BUCARAMANGA</td>\n",
       "      <td>76417311</td>\n",
       "      <td>...</td>\n",
       "      <td>10928</td>\n",
       "      <td>1297810</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10191146050</td>\n",
       "      <td>4</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2019-07-22 00:00:00</td>\n",
       "      <td>ESTUPIÑAN LOPEZ   ARMANDO</td>\n",
       "      <td>BUCARAMANGA</td>\n",
       "      <td>76710070</td>\n",
       "      <td>...</td>\n",
       "      <td>20805</td>\n",
       "      <td>1987240</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10191150337</td>\n",
       "      <td>5</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2020-07-03 00:00:00</td>\n",
       "      <td>OSMA   MARTHA ADELINA</td>\n",
       "      <td>BUCARAMANGA</td>\n",
       "      <td>76732213</td>\n",
       "      <td>...</td>\n",
       "      <td>66677</td>\n",
       "      <td>4464796</td>\n",
       "      <td>2125</td>\n",
       "      <td>0</td>\n",
       "      <td>340</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 41 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    obligacion  Titular  agencia       tipo primera_asignacion  \\\n",
       "0    201917264        1      201  MANTENIDO         2016-10-01   \n",
       "1  10191142053        2      101  MANTENIDO         2016-10-01   \n",
       "2  10191144858        3      101  MANTENIDO         2016-10-01   \n",
       "3  10191146050        4      101  MANTENIDO         2016-10-01   \n",
       "4  10191150337        5      101  MANTENIDO         2016-10-01   \n",
       "\n",
       "  fecha_asignacion  fecha_desasignacion  \\\n",
       "0       2016-10-01  2019-07-22 00:00:00   \n",
       "1       2016-10-01  2019-07-22 00:00:00   \n",
       "2       2016-10-01  2019-07-22 00:00:00   \n",
       "3       2016-10-01  2019-07-22 00:00:00   \n",
       "4       2016-10-01  2020-07-03 00:00:00   \n",
       "\n",
       "                                  nombre       ciudad  tel_casa  ...  \\\n",
       "0  SALAZAR ECHAVARRIA   CARMEN ADRIANA         CUCUTA  75702725  ...   \n",
       "1           GONZALEZ GONZALEZ   EFRAIN    BUCARAMANGA  76731342  ...   \n",
       "2        LANDAZABAL SANTANDER   MILENA    BUCARAMANGA  76417311  ...   \n",
       "3            ESTUPIÑAN LOPEZ   ARMANDO    BUCARAMANGA  76710070  ...   \n",
       "4                OSMA   MARTHA ADELINA    BUCARAMANGA  76732213  ...   \n",
       "\n",
       "   inter_corri  inter_mora   hyc seguro  iva  valor_cuota  \\\n",
       "0        29994     2397521     0      0    0            0   \n",
       "1            0      118354     0      0    0            0   \n",
       "2        10928     1297810     0      0    0            0   \n",
       "3        20805     1987240     0      0    0            0   \n",
       "4        66677     4464796  2125      0  340            0   \n",
       "\n",
       "   nro_cuotas_atrasadas cuotas_canceladas  cuotas_vencidas  vigente  \n",
       "0                     8                 2                6        0  \n",
       "1                    12                11                1        0  \n",
       "2                     8                 4                4        0  \n",
       "3                     8                 3                5        0  \n",
       "4                    14                 6                8        0  \n",
       "\n",
       "[5 rows x 41 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_asignacion_fdem =pd.read_excel(f'datos_avance.xlsx',sheet_name='asignacion_fdem')\n",
    "df_asignacion_fdem.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "6a7f87e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_asignacion_fdem.rename({'obligacion':'Credito'},axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "8ac7c709",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Credito', 'Titular', 'agencia', 'tipo', 'primera_asignacion',\n",
       "       'fecha_asignacion', 'fecha_desasignacion', 'nombre', 'ciudad',\n",
       "       'tel_casa', 'tel_negocio', 'celular', 'cedula_conyuge',\n",
       "       'nombre_conyuge', 'tel_casa_conyuge', 'tel_negocio_conyuge',\n",
       "       'celular_conyuge', 'nombre_garante', 'cedula_garante',\n",
       "       'tel_casa_garante', 'tel_negocio_garante', 'celular_garante',\n",
       "       'producto', 'dias_atraso', 'fecha_atraso', 'fecha_desembolso',\n",
       "       'fecha_castigo', 'grupo_castigo', 'ultimo_pago', 'monto', 'capital',\n",
       "       'inter_corri', 'inter_mora', 'hyc', 'seguro', 'iva', 'valor_cuota',\n",
       "       'nro_cuotas_atrasadas', 'cuotas_canceladas', 'cuotas_vencidas',\n",
       "       'vigente'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_asignacion_fdem.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2840efa0",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 165200 entries, 0 to 165199\n",
      "Data columns (total 41 columns):\n",
      " #   Column                Non-Null Count   Dtype         \n",
      "---  ------                --------------   -----         \n",
      " 0   Credito               165200 non-null  int64         \n",
      " 1   Titular               165200 non-null  int64         \n",
      " 2   agencia               165200 non-null  int64         \n",
      " 3   tipo                  165200 non-null  object        \n",
      " 4   primera_asignacion    165200 non-null  datetime64[ns]\n",
      " 5   fecha_asignacion      165200 non-null  datetime64[ns]\n",
      " 6   fecha_desasignacion   165200 non-null  object        \n",
      " 7   nombre                165200 non-null  object        \n",
      " 8   ciudad                165200 non-null  object        \n",
      " 9   tel_casa              165200 non-null  int64         \n",
      " 10  tel_negocio           165200 non-null  int64         \n",
      " 11  celular               165200 non-null  int64         \n",
      " 12  cedula_conyuge        165200 non-null  int64         \n",
      " 13  nombre_conyuge        71967 non-null   object        \n",
      " 14  tel_casa_conyuge      165200 non-null  int64         \n",
      " 15  tel_negocio_conyuge   165200 non-null  int64         \n",
      " 16  celular_conyuge       165200 non-null  int64         \n",
      " 17  nombre_garante        55074 non-null   object        \n",
      " 18  cedula_garante        165200 non-null  int64         \n",
      " 19  tel_casa_garante      165200 non-null  int64         \n",
      " 20  tel_negocio_garante   165200 non-null  int64         \n",
      " 21  celular_garante       165200 non-null  int64         \n",
      " 22  producto              165194 non-null  object        \n",
      " 23  dias_atraso           165200 non-null  int64         \n",
      " 24  fecha_atraso          165200 non-null  object        \n",
      " 25  fecha_desembolso      165200 non-null  object        \n",
      " 26  fecha_castigo         165200 non-null  object        \n",
      " 27  grupo_castigo         165200 non-null  datetime64[ns]\n",
      " 28  ultimo_pago           165200 non-null  object        \n",
      " 29  monto                 165200 non-null  int64         \n",
      " 30  capital               165200 non-null  int64         \n",
      " 31  inter_corri           165200 non-null  int64         \n",
      " 32  inter_mora            165200 non-null  int64         \n",
      " 33  hyc                   165200 non-null  int64         \n",
      " 34  seguro                165200 non-null  int64         \n",
      " 35  iva                   165200 non-null  int64         \n",
      " 36  valor_cuota           165200 non-null  int64         \n",
      " 37  nro_cuotas_atrasadas  165200 non-null  int64         \n",
      " 38  cuotas_canceladas     165200 non-null  int64         \n",
      " 39  cuotas_vencidas       165200 non-null  int64         \n",
      " 40  vigente               165200 non-null  int64         \n",
      "dtypes: datetime64[ns](3), int64(27), object(11)\n",
      "memory usage: 51.7+ MB\n"
     ]
    }
   ],
   "source": [
    "df_asignacion_fdem.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "06b9ecb9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Credito                 165200\n",
       "Titular                 134933\n",
       "agencia                    258\n",
       "tipo                        21\n",
       "primera_asignacion          27\n",
       "fecha_asignacion            26\n",
       "fecha_desasignacion         37\n",
       "nombre                  134889\n",
       "ciudad                     917\n",
       "tel_casa                 27086\n",
       "tel_negocio              67789\n",
       "celular                 126751\n",
       "cedula_conyuge           59920\n",
       "nombre_conyuge           59913\n",
       "tel_casa_conyuge          9392\n",
       "tel_negocio_conyuge       8701\n",
       "celular_conyuge          53610\n",
       "nombre_garante           48559\n",
       "cedula_garante           48668\n",
       "tel_casa_garante          8717\n",
       "tel_negocio_garante      10224\n",
       "celular_garante          42677\n",
       "producto                   111\n",
       "dias_atraso               4043\n",
       "fecha_atraso              5758\n",
       "fecha_desembolso          4418\n",
       "fecha_castigo               73\n",
       "grupo_castigo               64\n",
       "ultimo_pago               4437\n",
       "monto                    36795\n",
       "capital                 125973\n",
       "inter_corri             105073\n",
       "inter_mora              154567\n",
       "hyc                      75621\n",
       "seguro                   20088\n",
       "iva                      38908\n",
       "valor_cuota              62115\n",
       "nro_cuotas_atrasadas        54\n",
       "cuotas_canceladas           50\n",
       "cuotas_vencidas             61\n",
       "vigente                      2\n",
       "dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_asignacion_fdem.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ee9955a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_asignacion_fdem.to_csv(\"asignacion_fdem.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 604,
   "id": "d89dd71d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# df_resumen_creditos =pd.read_excel(f'datos_avance.xlsx',sheet_name='resumen_creditos')\n",
    "# df_resumen_creditos.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "590294ea",
   "metadata": {},
   "source": [
    "#### Rate of Recovery (Test #1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ed3149a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pagos.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8299340e",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pagos.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9d7c7ee9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "88c09413",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False    101214\n",
       "Name: SaldoFindeMes, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pagos['SaldoFindeMes'].isna().value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "06acff86",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "isinstance(df_pagos, pd.DataFrame)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "64615c82",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pagos['SaldoFindeMes'] = df_pagos['SaldoFindeMes'].replace(0, np.nan)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "9c7d9a5d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False    101172\n",
       "True         42\n",
       "Name: SaldoFindeMes, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pagos['SaldoFindeMes'].isna().value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "a8acead6",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pagos_cleaned = df_pagos.dropna(subset = [\"SaldoFindeMes\"]) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c35aa807",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False    101172\n",
       "Name: SaldoFindeMes, dtype: int64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pagos_cleaned[\"SaldoFindeMes\"].isna().value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c2697252",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_769/3106164577.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df_pagos_cleaned['rate'] = df_pagos_cleaned['TotalCancelado']/df_pagos_cleaned['SaldoFindeMes']\n"
     ]
    }
   ],
   "source": [
    "df_pagos_cleaned['rate'] = df_pagos_cleaned['TotalCancelado']/df_pagos_cleaned['SaldoFindeMes']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "7f226c8d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MesCargue</th>\n",
       "      <th>Codigo</th>\n",
       "      <th>Agencia</th>\n",
       "      <th>Credito</th>\n",
       "      <th>Estado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "      <th>DiasMoraFindeMes</th>\n",
       "      <th>SaldoActual</th>\n",
       "      <th>DiasmoraActual</th>\n",
       "      <th>FechaValor</th>\n",
       "      <th>...</th>\n",
       "      <th>PAGOSOTROS</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>VRgoech</th>\n",
       "      <th>CONDCAPITAL</th>\n",
       "      <th>CONDINTERES</th>\n",
       "      <th>CONDMORA</th>\n",
       "      <th>DiasMoraPago</th>\n",
       "      <th>Reverso</th>\n",
       "      <th>Concepto</th>\n",
       "      <th>rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>101</td>\n",
       "      <td>BUCARAMANGA AGENCIA PRINCIPAL</td>\n",
       "      <td>101101002728</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>321631.0</td>\n",
       "      <td>2032</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-18</td>\n",
       "      <td>...</td>\n",
       "      <td>10556</td>\n",
       "      <td>332187</td>\n",
       "      <td>321631</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>594634</td>\n",
       "      <td>2050</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.032820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>101</td>\n",
       "      <td>BUCARAMANGA AGENCIA PRINCIPAL</td>\n",
       "      <td>101104101267</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>149159.0</td>\n",
       "      <td>2142</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-31</td>\n",
       "      <td>...</td>\n",
       "      <td>4182</td>\n",
       "      <td>153341</td>\n",
       "      <td>149159</td>\n",
       "      <td>0</td>\n",
       "      <td>3498</td>\n",
       "      <td>356515</td>\n",
       "      <td>2172</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.028037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>101</td>\n",
       "      <td>BUCARAMANGA AGENCIA PRINCIPAL</td>\n",
       "      <td>101131041847</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>384108.0</td>\n",
       "      <td>490</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-20</td>\n",
       "      <td>...</td>\n",
       "      <td>9799</td>\n",
       "      <td>393907</td>\n",
       "      <td>384108</td>\n",
       "      <td>0</td>\n",
       "      <td>20307</td>\n",
       "      <td>233846</td>\n",
       "      <td>510</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.025511</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>102</td>\n",
       "      <td>BUCARAMANGA AGENCIA ASTURIAS</td>\n",
       "      <td>102141028221</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>291780.0</td>\n",
       "      <td>706</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-21</td>\n",
       "      <td>...</td>\n",
       "      <td>20420</td>\n",
       "      <td>312200</td>\n",
       "      <td>291780</td>\n",
       "      <td>0</td>\n",
       "      <td>30678</td>\n",
       "      <td>226165</td>\n",
       "      <td>727</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.069984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-10</td>\n",
       "      <td>102</td>\n",
       "      <td>BUCARAMANGA AGENCIA ASTURIAS</td>\n",
       "      <td>102141028298</td>\n",
       "      <td>Castigado</td>\n",
       "      <td>1363613.0</td>\n",
       "      <td>803</td>\n",
       "      <td>1322611</td>\n",
       "      <td>803</td>\n",
       "      <td>2016-10-18</td>\n",
       "      <td>...</td>\n",
       "      <td>12549</td>\n",
       "      <td>100000</td>\n",
       "      <td>87451</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>52487</td>\n",
       "      <td>821</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.073335</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MesCargue  Codigo                        Agencia       Credito     Estado  \\\n",
       "0   2016-10     101  BUCARAMANGA AGENCIA PRINCIPAL  101101002728  Castigado   \n",
       "1   2016-10     101  BUCARAMANGA AGENCIA PRINCIPAL  101104101267  Castigado   \n",
       "2   2016-10     101  BUCARAMANGA AGENCIA PRINCIPAL  101131041847  Castigado   \n",
       "3   2016-10     102   BUCARAMANGA AGENCIA ASTURIAS  102141028221  Castigado   \n",
       "4   2016-10     102   BUCARAMANGA AGENCIA ASTURIAS  102141028298  Castigado   \n",
       "\n",
       "   SaldoFindeMes  DiasMoraFindeMes  SaldoActual  DiasmoraActual FechaValor  \\\n",
       "0       321631.0              2032            0               0 2016-10-18   \n",
       "1       149159.0              2142            0               0 2016-10-31   \n",
       "2       384108.0               490            0               0 2016-10-20   \n",
       "3       291780.0               706            0               0 2016-10-21   \n",
       "4      1363613.0               803      1322611             803 2016-10-18   \n",
       "\n",
       "   ... PAGOSOTROS TotalCancelado  VRgoech  CONDCAPITAL CONDINTERES  CONDMORA  \\\n",
       "0  ...      10556         332187   321631            0           0    594634   \n",
       "1  ...       4182         153341   149159            0        3498    356515   \n",
       "2  ...       9799         393907   384108            0       20307    233846   \n",
       "3  ...      20420         312200   291780            0       30678    226165   \n",
       "4  ...      12549         100000    87451            0           0     52487   \n",
       "\n",
       "   DiasMoraPago  Reverso  Concepto      rate  \n",
       "0          2050      NaN       NaN  1.032820  \n",
       "1          2172      NaN       NaN  1.028037  \n",
       "2           510      NaN       NaN  1.025511  \n",
       "3           727      NaN       NaN  1.069984  \n",
       "4           821      NaN       NaN  0.073335  \n",
       "\n",
       "[5 rows x 28 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pagos_cleaned.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "792eaa6e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.series.Series'>\n",
      "Int64Index: 101172 entries, 0 to 101213\n",
      "Series name: SaldoFindeMes\n",
      "Non-Null Count   Dtype  \n",
      "--------------   -----  \n",
      "101172 non-null  float64\n",
      "dtypes: float64(1)\n",
      "memory usage: 1.5 MB\n"
     ]
    }
   ],
   "source": [
    "df_pagos_cleaned['SaldoFindeMes'].info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "1b1f159c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    101172.000000\n",
       "mean          0.440709\n",
       "std           1.849641\n",
       "min           0.000002\n",
       "25%           0.103865\n",
       "50%           0.221439\n",
       "75%           0.591595\n",
       "max         518.687059\n",
       "Name: rate, dtype: float64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pagos_cleaned['rate'].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c1557b09",
   "metadata": {},
   "source": [
    "## Merging Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0cc61b0f",
   "metadata": {},
   "source": [
    "### Merging Sum of Total Cancelado with Asignacion FDEM"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c19dfa96",
   "metadata": {},
   "source": [
    "From the Pagos Dataset, group by loan Id and sum all the payments made to that loan. Then merge it with the Bank dataset. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "e2585693",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_total = df_pagos.groupby('Credito')['TotalCancelado'].sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "755db1ee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Credito\n",
       "10191138571    149818\n",
       "10191147413    181360\n",
       "10691417316    102205\n",
       "10691418568    245626\n",
       "10691419227    261656\n",
       "Name: SaldoFindeMes, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_saldo = df_pagos.groupby('Credito')['SaldoFindeMes'].max()\n",
    "df_saldo.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "76532057",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Credito\n",
       "10191138571    310000\n",
       "10191147413    506660\n",
       "10691417316    104829\n",
       "10691418568    256732\n",
       "10691419227    500000\n",
       "Name: TotalCancelado, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_total.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "79c247a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_merge2 = df_asignacion_fdem.merge(df_total, on='Credito', how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "904e25ea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>Titular</th>\n",
       "      <th>agencia</th>\n",
       "      <th>tipo</th>\n",
       "      <th>primera_asignacion</th>\n",
       "      <th>fecha_asignacion</th>\n",
       "      <th>fecha_desasignacion</th>\n",
       "      <th>nombre</th>\n",
       "      <th>ciudad</th>\n",
       "      <th>tel_casa</th>\n",
       "      <th>...</th>\n",
       "      <th>inter_mora</th>\n",
       "      <th>hyc</th>\n",
       "      <th>seguro</th>\n",
       "      <th>iva</th>\n",
       "      <th>valor_cuota</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "      <th>TotalCancelado</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>101104101421</td>\n",
       "      <td>95</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>\\N</td>\n",
       "      <td>ZAPATA JIMENEZ   NATHALI</td>\n",
       "      <td>BUCARAMANGA</td>\n",
       "      <td>3175220817</td>\n",
       "      <td>...</td>\n",
       "      <td>28001</td>\n",
       "      <td>0</td>\n",
       "      <td>69</td>\n",
       "      <td>0</td>\n",
       "      <td>79913</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>695000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101111013583</td>\n",
       "      <td>140</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>\\N</td>\n",
       "      <td>FUENTES   CARLOS  EDUARDO</td>\n",
       "      <td>FLORIDABLANCA</td>\n",
       "      <td>76190026</td>\n",
       "      <td>...</td>\n",
       "      <td>207345</td>\n",
       "      <td>22601</td>\n",
       "      <td>437</td>\n",
       "      <td>4292</td>\n",
       "      <td>91658</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>320000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>101111014610</td>\n",
       "      <td>153</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2020-07-03 00:00:00</td>\n",
       "      <td>GARCIA RODRIGUEZ FRANKLIN GEOVANNY</td>\n",
       "      <td>FLORIDABLANCA</td>\n",
       "      <td>76196665</td>\n",
       "      <td>...</td>\n",
       "      <td>1774888</td>\n",
       "      <td>0</td>\n",
       "      <td>622</td>\n",
       "      <td>0</td>\n",
       "      <td>99815</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1300000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>101111015376</td>\n",
       "      <td>161</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2019-08-31 00:00:00</td>\n",
       "      <td>GOMEZ  LUIS RAMON</td>\n",
       "      <td>BUCARAMANGA</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>3448115</td>\n",
       "      <td>71755</td>\n",
       "      <td>1828</td>\n",
       "      <td>13636</td>\n",
       "      <td>132996</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1886000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>101111016438</td>\n",
       "      <td>180</td>\n",
       "      <td>101</td>\n",
       "      <td>MANTENIDO</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>2020-07-03 00:00:00</td>\n",
       "      <td>GONZALEZ RONDON MARIA DOLORES</td>\n",
       "      <td>BUCARAMANGA</td>\n",
       "      <td>76701592</td>\n",
       "      <td>...</td>\n",
       "      <td>153754</td>\n",
       "      <td>36059</td>\n",
       "      <td>825</td>\n",
       "      <td>6852</td>\n",
       "      <td>174804</td>\n",
       "      <td>15</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Credito  Titular  agencia       tipo primera_asignacion  \\\n",
       "0  101104101421       95      101  MANTENIDO         2016-10-01   \n",
       "1  101111013583      140      101  MANTENIDO         2016-10-01   \n",
       "2  101111014610      153      101  MANTENIDO         2016-10-01   \n",
       "3  101111015376      161      101  MANTENIDO         2016-10-01   \n",
       "4  101111016438      180      101  MANTENIDO         2016-10-01   \n",
       "\n",
       "  fecha_asignacion  fecha_desasignacion                              nombre  \\\n",
       "0       2016-10-01                   \\N          ZAPATA JIMENEZ   NATHALI     \n",
       "1       2016-10-01                   \\N           FUENTES   CARLOS  EDUARDO   \n",
       "2       2016-10-01  2020-07-03 00:00:00  GARCIA RODRIGUEZ FRANKLIN GEOVANNY   \n",
       "3       2016-10-01  2019-08-31 00:00:00                   GOMEZ  LUIS RAMON   \n",
       "4       2016-10-01  2020-07-03 00:00:00       GONZALEZ RONDON MARIA DOLORES   \n",
       "\n",
       "          ciudad    tel_casa  ...  inter_mora    hyc  seguro    iva  \\\n",
       "0    BUCARAMANGA  3175220817  ...       28001      0      69      0   \n",
       "1  FLORIDABLANCA    76190026  ...      207345  22601     437   4292   \n",
       "2  FLORIDABLANCA    76196665  ...     1774888      0     622      0   \n",
       "3    BUCARAMANGA           0  ...     3448115  71755    1828  13636   \n",
       "4    BUCARAMANGA    76701592  ...      153754  36059     825   6852   \n",
       "\n",
       "   valor_cuota  nro_cuotas_atrasadas  cuotas_canceladas cuotas_vencidas  \\\n",
       "0        79913                    12                  9               3   \n",
       "1        91658                    24                  8               6   \n",
       "2        99815                     8                  2               6   \n",
       "3       132996                    14                  4              10   \n",
       "4       174804                    15                  9               6   \n",
       "\n",
       "   vigente  TotalCancelado  \n",
       "0        0          695000  \n",
       "1        1          320000  \n",
       "2        0         1300000  \n",
       "3        0         1886000  \n",
       "4        0          150000  \n",
       "\n",
       "[5 rows x 42 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_merge2.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 622,
   "id": "58dc4986",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_merge2.to_csv(\"merged.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 733,
   "id": "b99b4c9e",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_merge3 = df_merge2.merge(df_saldo, on='Credito', how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 735,
   "id": "2680921c",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_merge3.to_csv(\"merged3.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "df8b4a7c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 15289 entries, 0 to 15288\n",
      "Data columns (total 10 columns):\n",
      " #   Column                Non-Null Count  Dtype \n",
      "---  ------                --------------  ----- \n",
      " 0   Credito               15289 non-null  int64 \n",
      " 1   nombre_conyuge        6355 non-null   object\n",
      " 2   nombre_garante        6497 non-null   object\n",
      " 3   dias_atraso           15289 non-null  int64 \n",
      " 4   monto                 15289 non-null  int64 \n",
      " 5   nro_cuotas_atrasadas  15289 non-null  int64 \n",
      " 6   cuotas_canceladas     15289 non-null  int64 \n",
      " 7   cuotas_vencidas       15289 non-null  int64 \n",
      " 8   vigente               15289 non-null  int64 \n",
      " 9   TotalCancelado        15289 non-null  int64 \n",
      "dtypes: int64(8), object(2)\n",
      "memory usage: 1.3+ MB\n"
     ]
    }
   ],
   "source": [
    "df_features2 = df_merge2[['Credito','nombre_conyuge','nombre_garante', 'dias_atraso', 'monto', 'nro_cuotas_atrasadas', 'cuotas_canceladas', 'cuotas_vencidas','vigente','TotalCancelado']]\n",
    "df_features2.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "bced17ce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>nombre_conyuge</th>\n",
       "      <th>nombre_garante</th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>monto</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "      <th>TotalCancelado</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>101104101421</td>\n",
       "      <td>AMAYA VELASQUEZ    CESAR AUGUSTO</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3157</td>\n",
       "      <td>820000</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>695000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101111013583</td>\n",
       "      <td>GARCIA   SUAREZ DIANA  HELENA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2117</td>\n",
       "      <td>1038863</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>320000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>101111014610</td>\n",
       "      <td>NaN</td>\n",
       "      <td>GARCIA   RODRIGUEZ  ELIZABETH</td>\n",
       "      <td>2727</td>\n",
       "      <td>700000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1300000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>101111015376</td>\n",
       "      <td>NaN</td>\n",
       "      <td>JAIMES   LIZARAZO DIEGO</td>\n",
       "      <td>2635</td>\n",
       "      <td>1500000</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1886000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>101111016438</td>\n",
       "      <td>MORENO   DIAZ RAUL</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2449</td>\n",
       "      <td>2030000</td>\n",
       "      <td>15</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Credito                      nombre_conyuge  \\\n",
       "0  101104101421  AMAYA VELASQUEZ    CESAR AUGUSTO     \n",
       "1  101111013583       GARCIA   SUAREZ DIANA  HELENA   \n",
       "2  101111014610                                 NaN   \n",
       "3  101111015376                                 NaN   \n",
       "4  101111016438                MORENO   DIAZ RAUL     \n",
       "\n",
       "                  nombre_garante  dias_atraso    monto  nro_cuotas_atrasadas  \\\n",
       "0                            NaN         3157   820000                    12   \n",
       "1                            NaN         2117  1038863                    24   \n",
       "2  GARCIA   RODRIGUEZ  ELIZABETH         2727   700000                     8   \n",
       "3       JAIMES   LIZARAZO DIEGO          2635  1500000                    14   \n",
       "4                            NaN         2449  2030000                    15   \n",
       "\n",
       "   cuotas_canceladas  cuotas_vencidas  vigente  TotalCancelado  \n",
       "0                  9                3        0          695000  \n",
       "1                  8                6        1          320000  \n",
       "2                  2                6        0         1300000  \n",
       "3                  4               10        0         1886000  \n",
       "4                  9                6        0          150000  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features2.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "89d10a3e",
   "metadata": {},
   "source": [
    "###### RR = Total Cancelado / MAX(SaldoFindeMes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "1a61f5c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features3 = df_features2.merge(df_saldo, on='Credito', how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "0e0bc8a6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>nombre_conyuge</th>\n",
       "      <th>nombre_garante</th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>monto</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>101104101421</td>\n",
       "      <td>AMAYA VELASQUEZ    CESAR AUGUSTO</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3157</td>\n",
       "      <td>820000</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>695000</td>\n",
       "      <td>760541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101111013583</td>\n",
       "      <td>GARCIA   SUAREZ DIANA  HELENA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2117</td>\n",
       "      <td>1038863</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>320000</td>\n",
       "      <td>715299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>101111014610</td>\n",
       "      <td>NaN</td>\n",
       "      <td>GARCIA   RODRIGUEZ  ELIZABETH</td>\n",
       "      <td>2727</td>\n",
       "      <td>700000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1300000</td>\n",
       "      <td>540311</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>101111015376</td>\n",
       "      <td>NaN</td>\n",
       "      <td>JAIMES   LIZARAZO DIEGO</td>\n",
       "      <td>2635</td>\n",
       "      <td>1500000</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1886000</td>\n",
       "      <td>1133161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>101111016438</td>\n",
       "      <td>MORENO   DIAZ RAUL</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2449</td>\n",
       "      <td>2030000</td>\n",
       "      <td>15</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "      <td>928039</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Credito                      nombre_conyuge  \\\n",
       "0  101104101421  AMAYA VELASQUEZ    CESAR AUGUSTO     \n",
       "1  101111013583       GARCIA   SUAREZ DIANA  HELENA   \n",
       "2  101111014610                                 NaN   \n",
       "3  101111015376                                 NaN   \n",
       "4  101111016438                MORENO   DIAZ RAUL     \n",
       "\n",
       "                  nombre_garante  dias_atraso    monto  nro_cuotas_atrasadas  \\\n",
       "0                            NaN         3157   820000                    12   \n",
       "1                            NaN         2117  1038863                    24   \n",
       "2  GARCIA   RODRIGUEZ  ELIZABETH         2727   700000                     8   \n",
       "3       JAIMES   LIZARAZO DIEGO          2635  1500000                    14   \n",
       "4                            NaN         2449  2030000                    15   \n",
       "\n",
       "   cuotas_canceladas  cuotas_vencidas  vigente  TotalCancelado  SaldoFindeMes  \n",
       "0                  9                3        0          695000         760541  \n",
       "1                  8                6        1          320000         715299  \n",
       "2                  2                6        0         1300000         540311  \n",
       "3                  4               10        0         1886000        1133161  \n",
       "4                  9                6        0          150000         928039  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1b31cf62",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "6769125b",
   "metadata": {},
   "source": [
    "## Data Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0dee3f75",
   "metadata": {},
   "source": [
    "### Conyuge (Spouse)\n",
    "\n",
    "\n",
    "Determinar si usuarios tienen conyuge o en caso contrario calificarlos como solteros"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6398132f",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7136                                  NaN\n",
       "6181                                  NaN\n",
       "10241                                 NaN\n",
       "176                                   NaN\n",
       "8068                                  NaN\n",
       "654                                   NaN\n",
       "10188                                 NaN\n",
       "14177    PIZARRO  CASTAÑEDA ANGIE TATIANA\n",
       "11824                                 NaN\n",
       "9196                                  NaN\n",
       "Name: nombre_conyuge, dtype: object"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3['nombre_conyuge'].sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "3c59e474",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features3['nombre_conyuge'].replace(to_replace=r\"[\\w'-]+\", value='Casado', regex = True, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "c26b70cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features3['nombre_conyuge'].fillna('Soltero', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "e22ff440",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features3.loc[df_features3['nombre_conyuge'].str.contains('Casado'), 'nombre_conyuge'] = 'Casado'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "8469f996",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "11465     Casado\n",
       "15271    Soltero\n",
       "10377    Soltero\n",
       "4008      Casado\n",
       "7415      Casado\n",
       "643      Soltero\n",
       "9475      Casado\n",
       "14866    Soltero\n",
       "4704     Soltero\n",
       "8558      Casado\n",
       "Name: nombre_conyuge, dtype: object"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3['nombre_conyuge'].sample(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a2f3588c",
   "metadata": {},
   "source": [
    "### Garante (Guarantor)\n",
    "\n",
    "Determinar si usuarios tienen garante"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "6f38846f",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features3['nombre_garante'].replace(to_replace=r\"[\\w'-]+\", value='Si', regex = True, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "d1a2d480",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features3['nombre_garante'].fillna('No', inplace=True)\n",
    "df_features3.loc[df_features3['nombre_garante'].str.contains('Si'), 'nombre_garante'] = 'Si'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "388ee93a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8034     No\n",
       "5654     Si\n",
       "10382    No\n",
       "7524     No\n",
       "13452    No\n",
       "9113     Si\n",
       "7934     No\n",
       "12486    No\n",
       "7394     No\n",
       "14210    Si\n",
       "Name: nombre_garante, dtype: object"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3['nombre_garante'].sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "1197851f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>nombre_conyuge</th>\n",
       "      <th>nombre_garante</th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>monto</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>101104101421</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>3157</td>\n",
       "      <td>820000</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>695000</td>\n",
       "      <td>760541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101111013583</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2117</td>\n",
       "      <td>1038863</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>320000</td>\n",
       "      <td>715299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>101111014610</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>Si</td>\n",
       "      <td>2727</td>\n",
       "      <td>700000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1300000</td>\n",
       "      <td>540311</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>101111015376</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>Si</td>\n",
       "      <td>2635</td>\n",
       "      <td>1500000</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1886000</td>\n",
       "      <td>1133161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>101111016438</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2449</td>\n",
       "      <td>2030000</td>\n",
       "      <td>15</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "      <td>928039</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Credito nombre_conyuge nombre_garante  dias_atraso    monto  \\\n",
       "0  101104101421         Casado             No         3157   820000   \n",
       "1  101111013583         Casado             No         2117  1038863   \n",
       "2  101111014610        Soltero             Si         2727   700000   \n",
       "3  101111015376        Soltero             Si         2635  1500000   \n",
       "4  101111016438         Casado             No         2449  2030000   \n",
       "\n",
       "   nro_cuotas_atrasadas  cuotas_canceladas  cuotas_vencidas  vigente  \\\n",
       "0                    12                  9                3        0   \n",
       "1                    24                  8                6        1   \n",
       "2                     8                  2                6        0   \n",
       "3                    14                  4               10        0   \n",
       "4                    15                  9                6        0   \n",
       "\n",
       "   TotalCancelado  SaldoFindeMes  \n",
       "0          695000         760541  \n",
       "1          320000         715299  \n",
       "2         1300000         540311  \n",
       "3         1886000        1133161  \n",
       "4          150000         928039  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "31fb5616",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Credito                 15289\n",
       "nombre_conyuge              2\n",
       "nombre_garante              2\n",
       "dias_atraso              1932\n",
       "monto                    5937\n",
       "nro_cuotas_atrasadas       42\n",
       "cuotas_canceladas          47\n",
       "cuotas_vencidas            44\n",
       "vigente                     2\n",
       "TotalCancelado           6856\n",
       "SaldoFindeMes           14314\n",
       "dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3.nunique()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cdb87928",
   "metadata": {},
   "source": [
    "## Defining Target Variable - Rate of Recovery"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d2286d8",
   "metadata": {},
   "source": [
    "The recovery rate is our target variable. We compute it as the sum of recovery payments made by a given loan Id, divided by the outstanding balance measured at the time of purchase of the NPL.\n",
    "It represents the percentage of exposure that can be recovered from each borrower through the debt collection process.\n",
    "\n",
    "\n",
    "\n",
    "Tasa de Recuperación: el porcentaje de exposición que se puede recuperar de cada usuario a través del proceso de cobro de deudas.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "9d3b5923",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "df_features3['rate'] = df_features3['TotalCancelado'] / df_features3['SaldoFindeMes']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "996a0f99",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>nombre_conyuge</th>\n",
       "      <th>nombre_garante</th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>monto</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "      <th>rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>101104101421</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>3157</td>\n",
       "      <td>820000</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>695000</td>\n",
       "      <td>760541</td>\n",
       "      <td>0.913823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101111013583</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2117</td>\n",
       "      <td>1038863</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>320000</td>\n",
       "      <td>715299</td>\n",
       "      <td>0.447365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>101111014610</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>Si</td>\n",
       "      <td>2727</td>\n",
       "      <td>700000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1300000</td>\n",
       "      <td>540311</td>\n",
       "      <td>2.406022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>101111015376</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>Si</td>\n",
       "      <td>2635</td>\n",
       "      <td>1500000</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1886000</td>\n",
       "      <td>1133161</td>\n",
       "      <td>1.664371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>101111016438</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2449</td>\n",
       "      <td>2030000</td>\n",
       "      <td>15</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "      <td>928039</td>\n",
       "      <td>0.161631</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Credito nombre_conyuge nombre_garante  dias_atraso    monto  \\\n",
       "0  101104101421         Casado             No         3157   820000   \n",
       "1  101111013583         Casado             No         2117  1038863   \n",
       "2  101111014610        Soltero             Si         2727   700000   \n",
       "3  101111015376        Soltero             Si         2635  1500000   \n",
       "4  101111016438         Casado             No         2449  2030000   \n",
       "\n",
       "   nro_cuotas_atrasadas  cuotas_canceladas  cuotas_vencidas  vigente  \\\n",
       "0                    12                  9                3        0   \n",
       "1                    24                  8                6        1   \n",
       "2                     8                  2                6        0   \n",
       "3                    14                  4               10        0   \n",
       "4                    15                  9                6        0   \n",
       "\n",
       "   TotalCancelado  SaldoFindeMes      rate  \n",
       "0          695000         760541  0.913823  \n",
       "1          320000         715299  0.447365  \n",
       "2         1300000         540311  2.406022  \n",
       "3         1886000        1133161  1.664371  \n",
       "4          150000         928039  0.161631  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "628eb494",
   "metadata": {},
   "source": [
    "### Rate of Recovery - Outliers"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92b1e4a4",
   "metadata": {},
   "source": [
    "Discard outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 677,
   "id": "3c8d30ea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    15288.000000\n",
       "mean         0.981606\n",
       "std          0.717497\n",
       "min          0.003761\n",
       "25%          0.452159\n",
       "50%          1.096228\n",
       "75%          1.340028\n",
       "max         27.428931\n",
       "Name: rate, dtype: float64"
      ]
     },
     "execution_count": 677,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features3['rate'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 678,
   "id": "61653374",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 678,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_features3['rate'].plot(kind='box')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 682,
   "id": "e2375e3c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 13133 entries, 0 to 15288\n",
      "Data columns (total 12 columns):\n",
      " #   Column                Non-Null Count  Dtype  \n",
      "---  ------                --------------  -----  \n",
      " 0   Credito               13133 non-null  int64  \n",
      " 1   nombre_conyuge        13133 non-null  object \n",
      " 2   nombre_garante        13133 non-null  object \n",
      " 3   dias_atraso           13133 non-null  int64  \n",
      " 4   monto                 13133 non-null  int64  \n",
      " 5   nro_cuotas_atrasadas  13133 non-null  int64  \n",
      " 6   cuotas_canceladas     13133 non-null  int64  \n",
      " 7   cuotas_vencidas       13133 non-null  int64  \n",
      " 8   vigente               13133 non-null  int64  \n",
      " 9   TotalCancelado        13133 non-null  int64  \n",
      " 10  SaldoFindeMes         13133 non-null  float64\n",
      " 11  rate                  13133 non-null  float64\n",
      "dtypes: float64(2), int64(8), object(2)\n",
      "memory usage: 1.3+ MB\n"
     ]
    }
   ],
   "source": [
    "df_out3 = df_features3[df_features3['rate'] < 1.5]  \n",
    "df_out3.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 683,
   "id": "48f0fb68",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 683,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_out3['rate'].plot(kind='box')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1f8e532",
   "metadata": {},
   "source": [
    "## Exploratory Data Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 684,
   "id": "ec099838",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_769/783989240.py:2: FutureWarning:\n",
      "\n",
      "The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "dataplot = sns.heatmap(df_out3.corr(), cmap=\"YlGnBu\", annot=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 685,
   "id": "d82a7ef1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3157</td>\n",
       "      <td>0.913823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2117</td>\n",
       "      <td>0.447365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2449</td>\n",
       "      <td>0.161631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2148</td>\n",
       "      <td>0.462452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2028</td>\n",
       "      <td>0.892272</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   dias_atraso      rate\n",
       "0         3157  0.913823\n",
       "1         2117  0.447365\n",
       "4         2449  0.161631\n",
       "6         2148  0.462452\n",
       "7         2028  0.892272"
      ]
     },
     "execution_count": 685,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_atraso_rate = df_out3[['dias_atraso','rate']]\n",
    "df_atraso_rate.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 686,
   "id": "45e380cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Plot dias atraso vs rate\n",
    "plt.scatter(df_out3['cuotas_canceladas'], df_out3['rate'])\n",
    "\n",
    "# Labels\n",
    "plt.xlabel(\"cuotas canceladas\")\n",
    "plt.ylabel(\"rate\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 687,
   "id": "cb2e200d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    13133.000000\n",
       "mean        10.227823\n",
       "std          6.744964\n",
       "min          0.000000\n",
       "25%          5.000000\n",
       "50%          9.000000\n",
       "75%         15.000000\n",
       "max         47.000000\n",
       "Name: cuotas_canceladas, dtype: float64"
      ]
     },
     "execution_count": 687,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    " df_out3['cuotas_canceladas'].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "65c26240",
   "metadata": {},
   "source": [
    "## Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 688,
   "id": "11b66c7c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 13133 entries, 0 to 15288\n",
      "Data columns (total 12 columns):\n",
      " #   Column                Non-Null Count  Dtype  \n",
      "---  ------                --------------  -----  \n",
      " 0   Credito               13133 non-null  int64  \n",
      " 1   nombre_conyuge        13133 non-null  object \n",
      " 2   nombre_garante        13133 non-null  object \n",
      " 3   dias_atraso           13133 non-null  int64  \n",
      " 4   monto                 13133 non-null  int64  \n",
      " 5   nro_cuotas_atrasadas  13133 non-null  int64  \n",
      " 6   cuotas_canceladas     13133 non-null  int64  \n",
      " 7   cuotas_vencidas       13133 non-null  int64  \n",
      " 8   vigente               13133 non-null  int64  \n",
      " 9   TotalCancelado        13133 non-null  int64  \n",
      " 10  SaldoFindeMes         13133 non-null  float64\n",
      " 11  rate                  13133 non-null  float64\n",
      "dtypes: float64(2), int64(8), object(2)\n",
      "memory usage: 1.3+ MB\n"
     ]
    }
   ],
   "source": [
    "df_out3.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 689,
   "id": "1e3db061",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>nombre_conyuge</th>\n",
       "      <th>nombre_garante</th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>monto</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "      <th>rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>101104101421</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>3157</td>\n",
       "      <td>820000</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>695000</td>\n",
       "      <td>760541.0</td>\n",
       "      <td>0.913823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101111013583</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2117</td>\n",
       "      <td>1038863</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>320000</td>\n",
       "      <td>715299.0</td>\n",
       "      <td>0.447365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>101111016438</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2449</td>\n",
       "      <td>2030000</td>\n",
       "      <td>15</td>\n",
       "      <td>9</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>150000</td>\n",
       "      <td>928039.0</td>\n",
       "      <td>0.161631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>101121024440</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2148</td>\n",
       "      <td>1200000</td>\n",
       "      <td>12</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>109000</td>\n",
       "      <td>235700.0</td>\n",
       "      <td>0.462452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>101121025332</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>2028</td>\n",
       "      <td>1530000</td>\n",
       "      <td>14</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>245000</td>\n",
       "      <td>274580.0</td>\n",
       "      <td>0.892272</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Credito nombre_conyuge nombre_garante  dias_atraso    monto  \\\n",
       "0  101104101421         Casado             No         3157   820000   \n",
       "1  101111013583         Casado             No         2117  1038863   \n",
       "4  101111016438         Casado             No         2449  2030000   \n",
       "6  101121024440         Casado             No         2148  1200000   \n",
       "7  101121025332         Casado             No         2028  1530000   \n",
       "\n",
       "   nro_cuotas_atrasadas  cuotas_canceladas  cuotas_vencidas  vigente  \\\n",
       "0                    12                  9                3        0   \n",
       "1                    24                  8                6        1   \n",
       "4                    15                  9                6        0   \n",
       "6                    12                 10                2        0   \n",
       "7                    14                 13                1        0   \n",
       "\n",
       "   TotalCancelado  SaldoFindeMes      rate  \n",
       "0          695000       760541.0  0.913823  \n",
       "1          320000       715299.0  0.447365  \n",
       "4          150000       928039.0  0.161631  \n",
       "6          109000       235700.0  0.462452  \n",
       "7          245000       274580.0  0.892272  "
      ]
     },
     "execution_count": 689,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_out3.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 690,
   "id": "599685a5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Credito', 'nombre_conyuge', 'nombre_garante', 'dias_atraso', 'monto',\n",
       "       'nro_cuotas_atrasadas', 'cuotas_canceladas', 'cuotas_vencidas',\n",
       "       'vigente', 'TotalCancelado', 'SaldoFindeMes', 'rate'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 690,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_out3.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 691,
   "id": "482a104f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Credito                 13133\n",
       "nombre_conyuge              2\n",
       "nombre_garante              2\n",
       "dias_atraso              1829\n",
       "monto                    5283\n",
       "nro_cuotas_atrasadas       42\n",
       "cuotas_canceladas          47\n",
       "cuotas_vencidas            44\n",
       "vigente                     2\n",
       "TotalCancelado           5704\n",
       "SaldoFindeMes           12381\n",
       "rate                    12967\n",
       "dtype: int64"
      ]
     },
     "execution_count": 691,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_out3.nunique()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7fe0642",
   "metadata": {},
   "source": [
    "### Split Train and Test sets"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c90954e5",
   "metadata": {},
   "source": [
    "Define X and y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 692,
   "id": "bebb4253",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df_out3.drop(columns=['rate','TotalCancelado','SaldoFindeMes'])\n",
    "y = df_out3['rate']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 693,
   "id": "8cc90747",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=88)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 694,
   "id": "cc67e6fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9193, 9)"
      ]
     },
     "execution_count": 694,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a87671d",
   "metadata": {},
   "source": [
    "#### Numerical Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 695,
   "id": "c493308c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['dias_atraso', 'monto', 'nro_cuotas_atrasadas',\n",
       "       'cuotas_canceladas', 'cuotas_vencidas', 'vigente'], dtype=object)"
      ]
     },
     "execution_count": 695,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feat_pre = X_train.drop(columns='Credito')\n",
    "feat_numerical_names = feat_pre.select_dtypes(include=['int64','float64']).nunique()\n",
    "feat_numerical_names = feat_numerical_names.index.values\n",
    "feat_numerical_names"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "74beeb2e",
   "metadata": {},
   "source": [
    "#### Categorical Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 696,
   "id": "72295ee2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "nombre_conyuge    2\n",
       "nombre_garante    2\n",
       "dtype: int64"
      ]
     },
     "execution_count": 696,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feat_categorical_nunique = X_train.select_dtypes(include=['object']).nunique()\n",
    "feat_categorical_nunique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 697,
   "id": "93c8287b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['nombre_conyuge', 'nombre_garante'], dtype=object)"
      ]
     },
     "execution_count": 697,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features_cat = feat_categorical_nunique.index.values\n",
    "features_cat"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9573c030",
   "metadata": {},
   "source": [
    "### Baseline Pipe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 698,
   "id": "7d382706",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline, FeatureUnion, make_union, make_pipeline\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.preprocessing import RobustScaler\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn import set_config; set_config(display='diagram')\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 699,
   "id": "a31c3501",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-f96e6023-1372-4402-a103-3004a44e8cfa {color: black;background-color: white;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa pre{padding: 0;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-toggleable {background-color: white;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-estimator:hover {background-color: #d4ebff;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-item {z-index: 1;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-parallel::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-parallel-item {display: flex;flex-direction: column;position: relative;background-color: white;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-parallel-item:only-child::after {width: 0;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;position: relative;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-label label {font-family: monospace;font-weight: bold;background-color: white;display: inline-block;line-height: 1.2em;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-label-container {position: relative;z-index: 2;text-align: center;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-f96e6023-1372-4402-a103-3004a44e8cfa div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-f96e6023-1372-4402-a103-3004a44e8cfa\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler())]),\n",
       "                                 array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                (&#x27;cat_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                 array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])</pre><b>Please rerun this cell to show the HTML repr or trust the notebook.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"44ee2449-e663-4bee-8d76-8a839b7974ad\" type=\"checkbox\" ><label for=\"44ee2449-e663-4bee-8d76-8a839b7974ad\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">ColumnTransformer</label><div class=\"sk-toggleable__content\"><pre>ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler())]),\n",
       "                                 array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                (&#x27;cat_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                 array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"062bf1ff-8916-4cff-9cec-36cf3cc95700\" type=\"checkbox\" ><label for=\"062bf1ff-8916-4cff-9cec-36cf3cc95700\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">num_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;dias_atraso&#x27; &#x27;monto&#x27; &#x27;nro_cuotas_atrasadas&#x27; &#x27;cuotas_canceladas&#x27;\n",
       " &#x27;cuotas_vencidas&#x27; &#x27;vigente&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"e9863be7-3f2f-4823-8fb1-c31623533e17\" type=\"checkbox\" ><label for=\"e9863be7-3f2f-4823-8fb1-c31623533e17\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">MinMaxScaler</label><div class=\"sk-toggleable__content\"><pre>MinMaxScaler()</pre></div></div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"4360efdf-5832-4cf0-b6c7-63ac14ec9b21\" type=\"checkbox\" ><label for=\"4360efdf-5832-4cf0-b6c7-63ac14ec9b21\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">cat_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;nombre_conyuge&#x27; &#x27;nombre_garante&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"8016d282-b570-40b7-8b54-55f76fd9f4b0\" type=\"checkbox\" ><label for=\"8016d282-b570-40b7-8b54-55f76fd9f4b0\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">OneHotEncoder</label><div class=\"sk-toggleable__content\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div></div></div></div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "ColumnTransformer(transformers=[('num_tr',\n",
       "                                 Pipeline(steps=[('scaler', MinMaxScaler())]),\n",
       "                                 array(['dias_atraso', 'monto', 'nro_cuotas_atrasadas',\n",
       "       'cuotas_canceladas', 'cuotas_vencidas', 'vigente'], dtype=object)),\n",
       "                                ('cat_tr',\n",
       "                                 Pipeline(steps=[('encoder',\n",
       "                                                  OneHotEncoder(handle_unknown='ignore'))]),\n",
       "                                 array(['nombre_conyuge', 'nombre_garante'], dtype=object))])"
      ]
     },
     "execution_count": 699,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Encode categorical values\n",
    "cat_transformer = Pipeline([\n",
    "    ('encoder', OneHotEncoder(handle_unknown='ignore'))\n",
    "     ])\n",
    "\n",
    "# Scale numerical values: \n",
    "num_transformer = Pipeline([\n",
    "    ('scaler', MinMaxScaler()),\n",
    "])\n",
    "\n",
    "# Parallelize \"num_transformer\" and \"cat_transfomer\"\n",
    "prepoc_baseline = ColumnTransformer([\n",
    "    ('num_tr', num_transformer, feat_numerical_names),\n",
    "    ('cat_tr', cat_transformer, features_cat)],\n",
    "    remainder='drop'\n",
    ")\n",
    "\n",
    "prepoc_baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 700,
   "id": "b66c8aa0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.03418958, 0.11492832, 0.4       , ..., 0.        , 1.        ,\n",
       "        0.        ],\n",
       "       [0.2301013 , 0.04196982, 0.4       , ..., 0.        , 0.        ,\n",
       "        1.        ],\n",
       "       [0.06078148, 0.08907755, 0.1       , ..., 1.        , 0.        ,\n",
       "        1.        ],\n",
       "       ...,\n",
       "       [0.0524602 , 0.06138087, 0.3       , ..., 0.        , 1.        ,\n",
       "        0.        ],\n",
       "       [0.15502894, 0.03349453, 0.26666667, ..., 0.        , 1.        ,\n",
       "        0.        ],\n",
       "       [0.00434153, 0.09849344, 0.4       , ..., 1.        , 0.        ,\n",
       "        1.        ]])"
      ]
     },
     "execution_count": 700,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_transformed = prepoc_baseline.fit_transform(X_train)\n",
    "X_transformed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 701,
   "id": "38a5a11c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.034190</td>\n",
       "      <td>0.114928</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.361702</td>\n",
       "      <td>0.145833</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.230101</td>\n",
       "      <td>0.041970</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.148936</td>\n",
       "      <td>0.354167</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.060781</td>\n",
       "      <td>0.089078</td>\n",
       "      <td>0.100000</td>\n",
       "      <td>0.042553</td>\n",
       "      <td>0.062500</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.178003</td>\n",
       "      <td>0.128403</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.208333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.073263</td>\n",
       "      <td>0.169144</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.170213</td>\n",
       "      <td>0.083333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9188</th>\n",
       "      <td>0.286360</td>\n",
       "      <td>0.029538</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.170213</td>\n",
       "      <td>0.083333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9189</th>\n",
       "      <td>0.143813</td>\n",
       "      <td>0.072776</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.382979</td>\n",
       "      <td>0.125000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9190</th>\n",
       "      <td>0.052460</td>\n",
       "      <td>0.061381</td>\n",
       "      <td>0.300000</td>\n",
       "      <td>0.106383</td>\n",
       "      <td>0.208333</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9191</th>\n",
       "      <td>0.155029</td>\n",
       "      <td>0.033495</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.276596</td>\n",
       "      <td>0.062500</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9192</th>\n",
       "      <td>0.004342</td>\n",
       "      <td>0.098493</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.191489</td>\n",
       "      <td>0.020833</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9193 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             0         1         2         3         4    5    6    7    8  \\\n",
       "0     0.034190  0.114928  0.400000  0.361702  0.145833  0.0  1.0  0.0  1.0   \n",
       "1     0.230101  0.041970  0.400000  0.148936  0.354167  0.0  1.0  0.0  0.0   \n",
       "2     0.060781  0.089078  0.100000  0.042553  0.062500  1.0  0.0  1.0  0.0   \n",
       "3     0.178003  0.128403  0.166667  0.000000  0.208333  0.0  0.0  1.0  0.0   \n",
       "4     0.073263  0.169144  0.200000  0.170213  0.083333  0.0  1.0  0.0  1.0   \n",
       "...        ...       ...       ...       ...       ...  ...  ...  ...  ...   \n",
       "9188  0.286360  0.029538  0.200000  0.170213  0.083333  0.0  0.0  1.0  0.0   \n",
       "9189  0.143813  0.072776  0.400000  0.382979  0.125000  0.0  0.0  1.0  1.0   \n",
       "9190  0.052460  0.061381  0.300000  0.106383  0.208333  1.0  1.0  0.0  1.0   \n",
       "9191  0.155029  0.033495  0.266667  0.276596  0.062500  0.0  1.0  0.0  1.0   \n",
       "9192  0.004342  0.098493  0.400000  0.191489  0.020833  1.0  0.0  1.0  0.0   \n",
       "\n",
       "        9  \n",
       "0     0.0  \n",
       "1     1.0  \n",
       "2     1.0  \n",
       "3     1.0  \n",
       "4     0.0  \n",
       "...   ...  \n",
       "9188  1.0  \n",
       "9189  0.0  \n",
       "9190  0.0  \n",
       "9191  0.0  \n",
       "9192  1.0  \n",
       "\n",
       "[9193 rows x 10 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(pd.DataFrame(X_transformed))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 702,
   "id": "69142238",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9193, 10)"
      ]
     },
     "execution_count": 702,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "shape_preproc_baseline = X_transformed.shape\n",
    "shape_preproc_baseline"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9864642a",
   "metadata": {},
   "source": [
    "### Adding estimators"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ec17752",
   "metadata": {},
   "source": [
    "Based on the literature review (https://www.researchgate.net/publication/342467001_Forecasting_recovery_rates_on_non-performing_loans_with_machine_learning)\n",
    "        \n",
    "We will try using the 3 best estimators for the task at hand and test them to see which one gives better results."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "806a1165",
   "metadata": {},
   "source": [
    "#### Gradient Boosting Regressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 703,
   "id": "7aabfc8f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 {color: black;background-color: white;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 pre{padding: 0;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-toggleable {background-color: white;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-estimator:hover {background-color: #d4ebff;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-item {z-index: 1;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-parallel::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-parallel-item {display: flex;flex-direction: column;position: relative;background-color: white;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-parallel-item:only-child::after {width: 0;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;position: relative;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-label label {font-family: monospace;font-weight: bold;background-color: white;display: inline-block;line-height: 1.2em;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-label-container {position: relative;z-index: 2;text-align: center;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-ebcbc2db-903b-4495-b30d-b35c17aa8001 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-ebcbc2db-903b-4495-b30d-b35c17aa8001\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;scaler&#x27;,\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                                 (&#x27;cat_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])),\n",
       "                (&#x27;gradientboostingregressor&#x27;, GradientBoostingRegressor())])</pre><b>Please rerun this cell to show the HTML repr or trust the notebook.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"2df49c1d-9c63-4087-9e8e-0c7d47293681\" type=\"checkbox\" ><label for=\"2df49c1d-9c63-4087-9e8e-0c7d47293681\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;scaler&#x27;,\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                                 (&#x27;cat_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])),\n",
       "                (&#x27;gradientboostingregressor&#x27;, GradientBoostingRegressor())])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"b330d27d-23cb-4e06-a47e-dc359b68452c\" type=\"checkbox\" ><label for=\"b330d27d-23cb-4e06-a47e-dc359b68452c\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">columntransformer: ColumnTransformer</label><div class=\"sk-toggleable__content\"><pre>ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler())]),\n",
       "                                 array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                (&#x27;cat_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                 array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"0d2adf22-6c34-40a5-984a-2f15ed05f4fd\" type=\"checkbox\" ><label for=\"0d2adf22-6c34-40a5-984a-2f15ed05f4fd\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">num_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;dias_atraso&#x27; &#x27;monto&#x27; &#x27;nro_cuotas_atrasadas&#x27; &#x27;cuotas_canceladas&#x27;\n",
       " &#x27;cuotas_vencidas&#x27; &#x27;vigente&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"95e1455f-2b87-4c22-99a8-5ab50638d514\" type=\"checkbox\" ><label for=\"95e1455f-2b87-4c22-99a8-5ab50638d514\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">MinMaxScaler</label><div class=\"sk-toggleable__content\"><pre>MinMaxScaler()</pre></div></div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"7097857b-19bc-4b52-9899-d9e56227bb9b\" type=\"checkbox\" ><label for=\"7097857b-19bc-4b52-9899-d9e56227bb9b\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">cat_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;nombre_conyuge&#x27; &#x27;nombre_garante&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"a619d895-0928-4d44-9291-46364d546a90\" type=\"checkbox\" ><label for=\"a619d895-0928-4d44-9291-46364d546a90\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">OneHotEncoder</label><div class=\"sk-toggleable__content\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div></div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"d1a6c02e-a114-41d6-b9e7-86dae403aeb8\" type=\"checkbox\" ><label for=\"d1a6c02e-a114-41d6-b9e7-86dae403aeb8\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">GradientBoostingRegressor</label><div class=\"sk-toggleable__content\"><pre>GradientBoostingRegressor()</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('columntransformer',\n",
       "                 ColumnTransformer(transformers=[('num_tr',\n",
       "                                                  Pipeline(steps=[('scaler',\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array(['dias_atraso', 'monto', 'nro_cuotas_atrasadas',\n",
       "       'cuotas_canceladas', 'cuotas_vencidas', 'vigente'], dtype=object)),\n",
       "                                                 ('cat_tr',\n",
       "                                                  Pipeline(steps=[('encoder',\n",
       "                                                                   OneHotEncoder(handle_unknown='ignore'))]),\n",
       "                                                  array(['nombre_conyuge', 'nombre_garante'], dtype=object))])),\n",
       "                ('gradientboostingregressor', GradientBoostingRegressor())])"
      ]
     },
     "execution_count": 703,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "\n",
    "\n",
    "pipe_baseline = make_pipeline(prepoc_baseline, GradientBoostingRegressor(n_estimators=100, verbose=0))\n",
    "\n",
    "pipe_baseline"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d5a06ee7",
   "metadata": {},
   "source": [
    "#### Extreme Gradient Tree Boosting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 704,
   "id": "b5b68bfd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a {color: black;background-color: white;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a pre{padding: 0;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-toggleable {background-color: white;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-estimator:hover {background-color: #d4ebff;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-item {z-index: 1;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-parallel::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-parallel-item {display: flex;flex-direction: column;position: relative;background-color: white;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-parallel-item:only-child::after {width: 0;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;position: relative;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-label label {font-family: monospace;font-weight: bold;background-color: white;display: inline-block;line-height: 1.2em;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-label-container {position: relative;z-index: 2;text-align: center;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-9d9af1f2-1861-4a8d-a4aa-b6a9086c594a\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;scaler&#x27;,\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                                 (&#x27;cat_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_...\n",
       "                              interaction_constraints=None, learning_rate=0.1,\n",
       "                              max_delta_step=None, max_depth=10,\n",
       "                              min_child_weight=None, missing=nan,\n",
       "                              monotone_constraints=None, n_estimators=100,\n",
       "                              n_jobs=None, num_parallel_tree=None,\n",
       "                              predictor=None, random_state=None, reg_alpha=None,\n",
       "                              reg_lambda=None, scale_pos_weight=None,\n",
       "                              subsample=None, tree_method=None,\n",
       "                              validate_parameters=None, verbosity=None))])</pre><b>Please rerun this cell to show the HTML repr or trust the notebook.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"317a644a-6542-4ece-a767-a7fe15978315\" type=\"checkbox\" ><label for=\"317a644a-6542-4ece-a767-a7fe15978315\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;scaler&#x27;,\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                                 (&#x27;cat_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_...\n",
       "                              interaction_constraints=None, learning_rate=0.1,\n",
       "                              max_delta_step=None, max_depth=10,\n",
       "                              min_child_weight=None, missing=nan,\n",
       "                              monotone_constraints=None, n_estimators=100,\n",
       "                              n_jobs=None, num_parallel_tree=None,\n",
       "                              predictor=None, random_state=None, reg_alpha=None,\n",
       "                              reg_lambda=None, scale_pos_weight=None,\n",
       "                              subsample=None, tree_method=None,\n",
       "                              validate_parameters=None, verbosity=None))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"180c071f-3596-41f4-bb61-6d7cf0ae0861\" type=\"checkbox\" ><label for=\"180c071f-3596-41f4-bb61-6d7cf0ae0861\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">columntransformer: ColumnTransformer</label><div class=\"sk-toggleable__content\"><pre>ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler())]),\n",
       "                                 array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                (&#x27;cat_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                 array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"a48cdf35-2d6c-43d9-81e3-9e14b496a853\" type=\"checkbox\" ><label for=\"a48cdf35-2d6c-43d9-81e3-9e14b496a853\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">num_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;dias_atraso&#x27; &#x27;monto&#x27; &#x27;nro_cuotas_atrasadas&#x27; &#x27;cuotas_canceladas&#x27;\n",
       " &#x27;cuotas_vencidas&#x27; &#x27;vigente&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"9c047b30-f3f7-40ab-acb6-6b088812a107\" type=\"checkbox\" ><label for=\"9c047b30-f3f7-40ab-acb6-6b088812a107\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">MinMaxScaler</label><div class=\"sk-toggleable__content\"><pre>MinMaxScaler()</pre></div></div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"ecec5ab8-563f-49c9-98b1-027c0e519098\" type=\"checkbox\" ><label for=\"ecec5ab8-563f-49c9-98b1-027c0e519098\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">cat_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;nombre_conyuge&#x27; &#x27;nombre_garante&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"5df1b79d-a819-4e56-ae16-471c3205dfc3\" type=\"checkbox\" ><label for=\"5df1b79d-a819-4e56-ae16-471c3205dfc3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">OneHotEncoder</label><div class=\"sk-toggleable__content\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div></div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"4f6c67c3-ed21-4884-ad58-312c6cf38ac4\" type=\"checkbox\" ><label for=\"4f6c67c3-ed21-4884-ad58-312c6cf38ac4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">XGBRegressor</label><div class=\"sk-toggleable__content\"><pre>XGBRegressor(base_score=None, booster=None, colsample_bylevel=None,\n",
       "             colsample_bynode=None, colsample_bytree=None,\n",
       "             enable_categorical=False, gamma=None, gpu_id=None,\n",
       "             importance_type=None, interaction_constraints=None,\n",
       "             learning_rate=0.1, max_delta_step=None, max_depth=10,\n",
       "             min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "             n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "             predictor=None, random_state=None, reg_alpha=None, reg_lambda=None,\n",
       "             scale_pos_weight=None, subsample=None, tree_method=None,\n",
       "             validate_parameters=None, verbosity=None)</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('columntransformer',\n",
       "                 ColumnTransformer(transformers=[('num_tr',\n",
       "                                                  Pipeline(steps=[('scaler',\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array(['dias_atraso', 'monto', 'nro_cuotas_atrasadas',\n",
       "       'cuotas_canceladas', 'cuotas_vencidas', 'vigente'], dtype=object)),\n",
       "                                                 ('cat_tr',\n",
       "                                                  Pipeline(steps=[('encoder',\n",
       "                                                                   OneHotEncoder(handle_unknown='ignore'))]),\n",
       "                                                  array(['nombre_conyuge', 'nombre_...\n",
       "                              interaction_constraints=None, learning_rate=0.1,\n",
       "                              max_delta_step=None, max_depth=10,\n",
       "                              min_child_weight=None, missing=nan,\n",
       "                              monotone_constraints=None, n_estimators=100,\n",
       "                              n_jobs=None, num_parallel_tree=None,\n",
       "                              predictor=None, random_state=None, reg_alpha=None,\n",
       "                              reg_lambda=None, scale_pos_weight=None,\n",
       "                              subsample=None, tree_method=None,\n",
       "                              validate_parameters=None, verbosity=None))])"
      ]
     },
     "execution_count": 704,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from xgboost import XGBRegressor\n",
    "\n",
    "pipe_baseline2 = make_pipeline(prepoc_baseline, XGBRegressor(max_depth=10, n_estimators=100, learning_rate=0.1))\n",
    "pipe_baseline2\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "61d57fa7",
   "metadata": {},
   "source": [
    "#### Random Forest Regressor "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 705,
   "id": "4151e61d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e {color: black;background-color: white;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e pre{padding: 0;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-toggleable {background-color: white;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-estimator:hover {background-color: #d4ebff;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-item {z-index: 1;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-parallel::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-parallel-item {display: flex;flex-direction: column;position: relative;background-color: white;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-parallel-item:only-child::after {width: 0;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;position: relative;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-label label {font-family: monospace;font-weight: bold;background-color: white;display: inline-block;line-height: 1.2em;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-label-container {position: relative;z-index: 2;text-align: center;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-db02b6d3-12ca-47ff-ab8c-be666e25e71e\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;scaler&#x27;,\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                                 (&#x27;cat_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])),\n",
       "                (&#x27;randomforestregressor&#x27;, RandomForestRegressor())])</pre><b>Please rerun this cell to show the HTML repr or trust the notebook.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"6a761c39-a740-4311-8f5b-ecfe8b077d3f\" type=\"checkbox\" ><label for=\"6a761c39-a740-4311-8f5b-ecfe8b077d3f\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;scaler&#x27;,\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                                 (&#x27;cat_tr&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])),\n",
       "                (&#x27;randomforestregressor&#x27;, RandomForestRegressor())])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"c5a60b54-4b68-4a89-b660-e2af381c7200\" type=\"checkbox\" ><label for=\"c5a60b54-4b68-4a89-b660-e2af381c7200\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">columntransformer: ColumnTransformer</label><div class=\"sk-toggleable__content\"><pre>ColumnTransformer(transformers=[(&#x27;num_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler())]),\n",
       "                                 array([&#x27;dias_atraso&#x27;, &#x27;monto&#x27;, &#x27;nro_cuotas_atrasadas&#x27;,\n",
       "       &#x27;cuotas_canceladas&#x27;, &#x27;cuotas_vencidas&#x27;, &#x27;vigente&#x27;], dtype=object)),\n",
       "                                (&#x27;cat_tr&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                 array([&#x27;nombre_conyuge&#x27;, &#x27;nombre_garante&#x27;], dtype=object))])</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"3c9f912c-a09d-4691-a08e-dfaa55deb644\" type=\"checkbox\" ><label for=\"3c9f912c-a09d-4691-a08e-dfaa55deb644\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">num_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;dias_atraso&#x27; &#x27;monto&#x27; &#x27;nro_cuotas_atrasadas&#x27; &#x27;cuotas_canceladas&#x27;\n",
       " &#x27;cuotas_vencidas&#x27; &#x27;vigente&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"f5f9f244-9176-40b3-b324-f2c6690a46b4\" type=\"checkbox\" ><label for=\"f5f9f244-9176-40b3-b324-f2c6690a46b4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">MinMaxScaler</label><div class=\"sk-toggleable__content\"><pre>MinMaxScaler()</pre></div></div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"9c424833-2580-44a0-8bce-42474f9eee24\" type=\"checkbox\" ><label for=\"9c424833-2580-44a0-8bce-42474f9eee24\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">cat_tr</label><div class=\"sk-toggleable__content\"><pre>[&#x27;nombre_conyuge&#x27; &#x27;nombre_garante&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"be9b7ca2-d1a2-4eed-a86e-306113a0e237\" type=\"checkbox\" ><label for=\"be9b7ca2-d1a2-4eed-a86e-306113a0e237\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">OneHotEncoder</label><div class=\"sk-toggleable__content\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div></div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"4d757288-2093-40e5-8f5f-a869b09551b9\" type=\"checkbox\" ><label for=\"4d757288-2093-40e5-8f5f-a869b09551b9\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomForestRegressor</label><div class=\"sk-toggleable__content\"><pre>RandomForestRegressor()</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('columntransformer',\n",
       "                 ColumnTransformer(transformers=[('num_tr',\n",
       "                                                  Pipeline(steps=[('scaler',\n",
       "                                                                   MinMaxScaler())]),\n",
       "                                                  array(['dias_atraso', 'monto', 'nro_cuotas_atrasadas',\n",
       "       'cuotas_canceladas', 'cuotas_vencidas', 'vigente'], dtype=object)),\n",
       "                                                 ('cat_tr',\n",
       "                                                  Pipeline(steps=[('encoder',\n",
       "                                                                   OneHotEncoder(handle_unknown='ignore'))]),\n",
       "                                                  array(['nombre_conyuge', 'nombre_garante'], dtype=object))])),\n",
       "                ('randomforestregressor', RandomForestRegressor())])"
      ]
     },
     "execution_count": 705,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "pipe_baseline3 = make_pipeline(prepoc_baseline, RandomForestRegressor(n_estimators=100))\n",
    "pipe_baseline3"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84b95b35",
   "metadata": {},
   "source": [
    "### Cross-Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 706,
   "id": "3cd37469",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "make_scorer(mean_squared_log_error, squared=False)"
      ]
     },
     "execution_count": 706,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error, make_scorer\n",
    "from sklearn.metrics import mean_squared_log_error\n",
    "\n",
    "\n",
    "rmsle = make_scorer(mean_squared_log_error,squared=False)\n",
    "rmsle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 707,
   "id": "b27adef5",
   "metadata": {},
   "outputs": [],
   "source": [
    "rmsle_neg = make_scorer(mean_squared_log_error,squared=False,greater_is_better=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 708,
   "id": "e01551e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_transformed =pd.DataFrame(X_transformed)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 709,
   "id": "8fb84b12",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.2650310469697699"
      ]
     },
     "execution_count": 709,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import cross_validate\n",
    "\n",
    "score_baseline = cross_val_score(pipe_baseline, X_train, y_train, cv=5, scoring=rmsle).mean()\n",
    "score_baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 710,
   "id": "505b2edb",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.2775116522451355"
      ]
     },
     "execution_count": 710,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_baseline2 = cross_val_score(pipe_baseline2, X_train, y_train, cv=5, scoring=rmsle).mean()\n",
    "score_baseline2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 711,
   "id": "3f14999a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.2747966505945331"
      ]
     },
     "execution_count": 711,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_baseline3 = cross_val_score(pipe_baseline3, X_train, y_train, cv=5, scoring=rmsle).mean()\n",
    "score_baseline3"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a282c24",
   "metadata": {},
   "source": [
    "Gradient Boosting proves better than Random Forest and XBGR. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7b5dcfd",
   "metadata": {},
   "source": [
    "#### Predict baseline\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 744,
   "id": "abd72eb1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>nombre_conyuge</th>\n",
       "      <th>nombre_garante</th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>monto</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14186</th>\n",
       "      <td>810171011686</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>287</td>\n",
       "      <td>1095000</td>\n",
       "      <td>15</td>\n",
       "      <td>6</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2754</th>\n",
       "      <td>203141042347</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>No</td>\n",
       "      <td>985</td>\n",
       "      <td>2200000</td>\n",
       "      <td>24</td>\n",
       "      <td>21</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10557</th>\n",
       "      <td>648161007515</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>533</td>\n",
       "      <td>1140060</td>\n",
       "      <td>14</td>\n",
       "      <td>11</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11029</th>\n",
       "      <td>658181006983</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>No</td>\n",
       "      <td>461</td>\n",
       "      <td>4267480</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1036</th>\n",
       "      <td>109181033263</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>167</td>\n",
       "      <td>1300000</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>887</th>\n",
       "      <td>107171071219</td>\n",
       "      <td>Casado</td>\n",
       "      <td>No</td>\n",
       "      <td>145</td>\n",
       "      <td>1150000</td>\n",
       "      <td>12</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9362</th>\n",
       "      <td>623171013332</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>No</td>\n",
       "      <td>205</td>\n",
       "      <td>1050000</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10483</th>\n",
       "      <td>647131001637</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>Si</td>\n",
       "      <td>1735</td>\n",
       "      <td>2057650</td>\n",
       "      <td>18</td>\n",
       "      <td>8</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5301</th>\n",
       "      <td>402141050665</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>No</td>\n",
       "      <td>1434</td>\n",
       "      <td>800000</td>\n",
       "      <td>10</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5683</th>\n",
       "      <td>404181020034</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>Si</td>\n",
       "      <td>198</td>\n",
       "      <td>13098912</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3940 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Credito nombre_conyuge nombre_garante  dias_atraso     monto  \\\n",
       "14186  810171011686         Casado             No          287   1095000   \n",
       "2754   203141042347        Soltero             No          985   2200000   \n",
       "10557  648161007515         Casado             No          533   1140060   \n",
       "11029  658181006983        Soltero             No          461   4267480   \n",
       "1036   109181033263         Casado             No          167   1300000   \n",
       "...             ...            ...            ...          ...       ...   \n",
       "887    107171071219         Casado             No          145   1150000   \n",
       "9362   623171013332        Soltero             No          205   1050000   \n",
       "10483  647131001637        Soltero             Si         1735   2057650   \n",
       "5301   402141050665        Soltero             No         1434    800000   \n",
       "5683   404181020034        Soltero             Si          198  13098912   \n",
       "\n",
       "       nro_cuotas_atrasadas  cuotas_canceladas  cuotas_vencidas  vigente  \n",
       "14186                    15                  6                9        0  \n",
       "2754                     24                 21                3        0  \n",
       "10557                    14                 11                3        0  \n",
       "11029                     8                  3                5        0  \n",
       "1036                      4                  2                2        0  \n",
       "...                     ...                ...              ...      ...  \n",
       "887                      12                 10                2        0  \n",
       "9362                     12                  9                3        0  \n",
       "10483                    18                  8               10        0  \n",
       "5301                     10                  6                4        0  \n",
       "5683                      3                  8                7        0  \n",
       "\n",
       "[3940 rows x 9 columns]"
      ]
     },
     "execution_count": 744,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 713,
   "id": "b7d99155",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.87946006, 0.84472249, 1.05426292, ..., 0.8300004 , 0.65136532,\n",
       "       0.32960041])"
      ]
     },
     "execution_count": 713,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe_baseline.fit(X_train,y_train)\n",
    "y_pred_baseline = pipe_baseline.predict(X_test)\n",
    "y_pred_baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 738,
   "id": "2b6b51f3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3940,)"
      ]
     },
     "execution_count": 738,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_baseline.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 715,
   "id": "7d5df50f",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test.Credito.reset_index(drop=True, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 716,
   "id": "bb1173fa",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>810171011686</td>\n",
       "      <td>0.879460</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>203141042347</td>\n",
       "      <td>0.844722</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>648161007515</td>\n",
       "      <td>1.054263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>658181006983</td>\n",
       "      <td>0.696153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>109181033263</td>\n",
       "      <td>0.959952</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3935</th>\n",
       "      <td>107171071219</td>\n",
       "      <td>1.081688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3936</th>\n",
       "      <td>623171013332</td>\n",
       "      <td>1.006937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3937</th>\n",
       "      <td>647131001637</td>\n",
       "      <td>0.830000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3938</th>\n",
       "      <td>402141050665</td>\n",
       "      <td>0.651365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3939</th>\n",
       "      <td>404181020034</td>\n",
       "      <td>0.329600</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3940 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           Credito      rate\n",
       "0     810171011686  0.879460\n",
       "1     203141042347  0.844722\n",
       "2     648161007515  1.054263\n",
       "3     658181006983  0.696153\n",
       "4     109181033263  0.959952\n",
       "...            ...       ...\n",
       "3935  107171071219  1.081688\n",
       "3936  623171013332  1.006937\n",
       "3937  647131001637  0.830000\n",
       "3938  402141050665  0.651365\n",
       "3939  404181020034  0.329600\n",
       "\n",
       "[3940 rows x 2 columns]"
      ]
     },
     "execution_count": 716,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission_baseline = pd.concat([X_test['Credito'], \n",
    "                     pd.Series(y_pred_baseline, \n",
    "                               name=\"rate\")], \n",
    "                    axis=1)\n",
    "submission_baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 717,
   "id": "bfe6d0f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_final = submission_baseline"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b01deb68",
   "metadata": {},
   "source": [
    "## Forecasting Rate of Recovery"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d92c4be",
   "metadata": {},
   "source": [
    "Sorted list by loan id with their rate of recovery forecast."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 725,
   "id": "cea98c51",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1054</th>\n",
       "      <td>107201083664</td>\n",
       "      <td>0.119286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2223</th>\n",
       "      <td>201201018267</td>\n",
       "      <td>0.215419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1333</th>\n",
       "      <td>201201014848</td>\n",
       "      <td>0.231080</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2577</th>\n",
       "      <td>208161052472</td>\n",
       "      <td>0.244925</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3574</th>\n",
       "      <td>220211011399</td>\n",
       "      <td>0.294589</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2210</th>\n",
       "      <td>302161097923</td>\n",
       "      <td>0.327799</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3939</th>\n",
       "      <td>404181020034</td>\n",
       "      <td>0.329600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1940</th>\n",
       "      <td>623201018376</td>\n",
       "      <td>0.343598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2547</th>\n",
       "      <td>610191013178</td>\n",
       "      <td>0.343854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>706</th>\n",
       "      <td>307201015147</td>\n",
       "      <td>0.347262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3320</th>\n",
       "      <td>107201082726</td>\n",
       "      <td>0.366290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2667</th>\n",
       "      <td>715201019032</td>\n",
       "      <td>0.366635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>615</th>\n",
       "      <td>610181010662</td>\n",
       "      <td>0.367847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2457</th>\n",
       "      <td>719181012818</td>\n",
       "      <td>0.394076</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2815</th>\n",
       "      <td>669201008252</td>\n",
       "      <td>0.400693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3297</th>\n",
       "      <td>827191007656</td>\n",
       "      <td>0.413701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>958</th>\n",
       "      <td>305191018856</td>\n",
       "      <td>0.413701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2302</th>\n",
       "      <td>502201051729</td>\n",
       "      <td>0.415776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2779</th>\n",
       "      <td>136171016230</td>\n",
       "      <td>0.422050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3158</th>\n",
       "      <td>650201009364</td>\n",
       "      <td>0.425670</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Credito      rate\n",
       "1054  107201083664  0.119286\n",
       "2223  201201018267  0.215419\n",
       "1333  201201014848  0.231080\n",
       "2577  208161052472  0.244925\n",
       "3574  220211011399  0.294589\n",
       "2210  302161097923  0.327799\n",
       "3939  404181020034  0.329600\n",
       "1940  623201018376  0.343598\n",
       "2547  610191013178  0.343854\n",
       "706   307201015147  0.347262\n",
       "3320  107201082726  0.366290\n",
       "2667  715201019032  0.366635\n",
       "615   610181010662  0.367847\n",
       "2457  719181012818  0.394076\n",
       "2815  669201008252  0.400693\n",
       "3297  827191007656  0.413701\n",
       "958   305191018856  0.413701\n",
       "2302  502201051729  0.415776\n",
       "2779  136171016230  0.422050\n",
       "3158  650201009364  0.425670"
      ]
     },
     "execution_count": 725,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sorted = df_final.sort_values(by=['rate'],ascending=True)\n",
    "df_sorted.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ad3d49e",
   "metadata": {},
   "source": [
    "### Loan Selector  (Buscador de Credito)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 743,
   "id": "f5bd5bc6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Credito</th>\n",
       "      <th>nombre_conyuge</th>\n",
       "      <th>nombre_garante</th>\n",
       "      <th>dias_atraso</th>\n",
       "      <th>monto</th>\n",
       "      <th>nro_cuotas_atrasadas</th>\n",
       "      <th>cuotas_canceladas</th>\n",
       "      <th>cuotas_vencidas</th>\n",
       "      <th>vigente</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "      <th>rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>914</th>\n",
       "      <td>107201083664</td>\n",
       "      <td>Soltero</td>\n",
       "      <td>No</td>\n",
       "      <td>440</td>\n",
       "      <td>12387022</td>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>797966</td>\n",
       "      <td>12387022.0</td>\n",
       "      <td>0.06442</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Credito nombre_conyuge nombre_garante  dias_atraso     monto  \\\n",
       "914  107201083664        Soltero             No          440  12387022   \n",
       "\n",
       "     nro_cuotas_atrasadas  cuotas_canceladas  cuotas_vencidas  vigente  \\\n",
       "914                    60                  0               15        0   \n",
       "\n",
       "     TotalCancelado  SaldoFindeMes     rate  \n",
       "914          797966     12387022.0  0.06442  "
      ]
     },
     "execution_count": 743,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_out3.loc[df_out3['Credito'] ==107201083664]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 727,
   "id": "ab050d9c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MesCargue</th>\n",
       "      <th>Codigo</th>\n",
       "      <th>Agencia</th>\n",
       "      <th>Credito</th>\n",
       "      <th>Estado</th>\n",
       "      <th>SaldoFindeMes</th>\n",
       "      <th>DiasMoraFindeMes</th>\n",
       "      <th>SaldoActual</th>\n",
       "      <th>DiasmoraActual</th>\n",
       "      <th>FechaValor</th>\n",
       "      <th>...</th>\n",
       "      <th>MORA</th>\n",
       "      <th>PAGOSOTROS</th>\n",
       "      <th>TotalCancelado</th>\n",
       "      <th>VRgoech</th>\n",
       "      <th>CONDCAPITAL</th>\n",
       "      <th>CONDINTERES</th>\n",
       "      <th>CONDMORA</th>\n",
       "      <th>DiasMoraPago</th>\n",
       "      <th>Reverso</th>\n",
       "      <th>Concepto</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>95608</th>\n",
       "      <td>2022-05</td>\n",
       "      <td>0</td>\n",
       "      <td>PIEDECUESTA CENTRO</td>\n",
       "      <td>107201083664</td>\n",
       "      <td>castigado</td>\n",
       "      <td>12387022.0</td>\n",
       "      <td>440</td>\n",
       "      <td>12329419</td>\n",
       "      <td>441</td>\n",
       "      <td>2022-05-02</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>54084</td>\n",
       "      <td>797966</td>\n",
       "      <td>743882</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>300957</td>\n",
       "      <td>442</td>\n",
       "      <td>NO</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      MesCargue  Codigo             Agencia       Credito     Estado  \\\n",
       "95608   2022-05       0  PIEDECUESTA CENTRO  107201083664  castigado   \n",
       "\n",
       "       SaldoFindeMes  DiasMoraFindeMes  SaldoActual  DiasmoraActual  \\\n",
       "95608     12387022.0               440     12329419             441   \n",
       "\n",
       "      FechaValor  ... MORA PAGOSOTROS  TotalCancelado  VRgoech CONDCAPITAL  \\\n",
       "95608 2022-05-02  ...    0      54084          797966   743882           0   \n",
       "\n",
       "       CONDINTERES  CONDMORA  DiasMoraPago  Reverso  Concepto  \n",
       "95608            0    300957           442       NO         0  \n",
       "\n",
       "[1 rows x 27 columns]"
      ]
     },
     "execution_count": 727,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pagos.loc[df_pagos['Credito'] ==107201083664]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de37d84b",
   "metadata": {},
   "source": [
    "#### Class Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 745,
   "id": "cbeb2dea",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np  \n",
    "import pandas as pd  \n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.utils.validation import check_is_fitted\n",
    "import plotly.express as px\n",
    "\n",
    "\n",
    "class FeatureImportance:\n",
    "\n",
    "    \"\"\"\n",
    "    \n",
    "    Extract & Plot the Feature Names & Importance Values from a Scikit-Learn Pipeline.\n",
    "    \n",
    "    The input is a Pipeline that starts with a ColumnTransformer & ends with a regression or classification model. \n",
    "    As intermediate steps, the Pipeline can have any number or no instances from sklearn.feature_selection.\n",
    "\n",
    "    Note: \n",
    "    If the ColumnTransformer contains Pipelines and if one of the transformers in the Pipeline is adding completely new columns, \n",
    "    it must come last in the pipeline. For example, OneHotEncoder, MissingIndicator & SimpleImputer(add_indicator=True) add columns \n",
    "    to the dataset that didn't exist before, so there should come last in the Pipeline.\n",
    "    \n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    pipeline : a Scikit-learn Pipeline class where the a ColumnTransformer is the first element and model estimator is the last element\n",
    "    verbose : a boolean. Whether to print all of the diagnostics. Default is False.\n",
    "    \n",
    "    Attributes\n",
    "    __________\n",
    "    column_transformer_features :  A list of the feature names created by the ColumnTransformer prior to any selectors being applied\n",
    "    transformer_list : A list of the transformer names that correspond with the `column_transformer_features` attribute\n",
    "    discarded_features : A list of the features names that were not selected by a sklearn.feature_selection instance.\n",
    "    discarding_selectors : A list of the selector names corresponding with the `discarded_features` attribute\n",
    "    feature_importance :  A Pandas Series containing the feature importance values and feature names as the index.    \n",
    "    plot_importances_df : A Pandas DataFrame containing the subset of features and values that are actually displaced in the plot. \n",
    "    feature_info_df : A Pandas DataFrame that aggregates the other attributes. The index is column_transformer_features. The transformer column contains the transformer_list.\n",
    "        value contains the feature_importance values. discarding_selector contains discarding_selectors & is_retained is a Boolean indicating whether the feature was retained.\n",
    "    \n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    def __init__(self, pipeline, verbose=False):\n",
    "        self.pipeline = pipeline\n",
    "        self.verbose = verbose\n",
    "\n",
    "\n",
    "    def get_feature_names(self, verbose=None):  \n",
    "\n",
    "        \"\"\"\n",
    "\n",
    "        Get the column names from the a ColumnTransformer containing transformers & pipelines\n",
    "\n",
    "        Parameters\n",
    "        ----------\n",
    "        verbose : a boolean indicating whether to print summaries. \n",
    "            default = False\n",
    "\n",
    "\n",
    "        Returns\n",
    "        -------\n",
    "        a list of the correct feature names\n",
    "\n",
    "        Note: \n",
    "        If the ColumnTransformer contains Pipelines and if one of the transformers in the Pipeline is adding completely new columns, \n",
    "        it must come last in the pipeline. For example, OneHotEncoder, MissingIndicator & SimpleImputer(add_indicator=True) add columns \n",
    "        to the dataset that didn't exist before, so there should come last in the Pipeline.\n",
    "\n",
    "        Inspiration: https://github.com/scikit-learn/scikit-learn/issues/12525 \n",
    "\n",
    "        \"\"\"\n",
    "\n",
    "        if verbose is None:\n",
    "            verbose = self.verbose\n",
    "            \n",
    "        if verbose: print('''\\n\\n---------\\nRunning get_feature_names\\n---------\\n''')\n",
    "        \n",
    "        column_transformer = self.pipeline[0]        \n",
    "        assert isinstance(column_transformer, ColumnTransformer), \"Input isn't a ColumnTransformer\"\n",
    "        check_is_fitted(column_transformer)\n",
    "\n",
    "        new_feature_names, transformer_list = [], []\n",
    "\n",
    "        for i, transformer_item in enumerate(column_transformer.transformers_): \n",
    "            \n",
    "            transformer_name, transformer, orig_feature_names = transformer_item\n",
    "            orig_feature_names = list(orig_feature_names)\n",
    "            \n",
    "            if verbose: \n",
    "                print('\\n\\n', i, '. Transformer/Pipeline: ', transformer_name, ',', \n",
    "                      transformer.__class__.__name__, '\\n')\n",
    "                print('\\tn_orig_feature_names:', len(orig_feature_names))\n",
    "\n",
    "            if transformer == 'drop':\n",
    "                    \n",
    "                continue\n",
    "                \n",
    "            if isinstance(transformer, Pipeline):\n",
    "                # if pipeline, get the last transformer in the Pipeline\n",
    "                transformer = transformer.steps[-1][1]\n",
    "\n",
    "            if hasattr(transformer, 'get_feature_names'):\n",
    "\n",
    "                if 'input_features' in transformer.get_feature_names.__code__.co_varnames:\n",
    "\n",
    "                    names = list(transformer.get_feature_names(orig_feature_names))\n",
    "\n",
    "                else:\n",
    "\n",
    "                    names = list(transformer.get_feature_names())\n",
    "\n",
    "            elif hasattr(transformer,'indicator_') and transformer.add_indicator:\n",
    "                # is this transformer one of the imputers & did it call the MissingIndicator?\n",
    "\n",
    "                missing_indicator_indices = transformer.indicator_.features_\n",
    "                missing_indicators = [orig_feature_names[idx] + '_missing_flag'\\\n",
    "                                      for idx in missing_indicator_indices]\n",
    "                names = orig_feature_names + missing_indicators\n",
    "\n",
    "            elif hasattr(transformer,'features_'):\n",
    "                # is this a MissingIndicator class? \n",
    "                missing_indicator_indices = transformer.features_\n",
    "                missing_indicators = [orig_feature_names[idx] + '_missing_flag'\\\n",
    "                                      for idx in missing_indicator_indices]\n",
    "\n",
    "            else:\n",
    "\n",
    "                names = orig_feature_names\n",
    "\n",
    "            if verbose: \n",
    "                print('\\tn_new_features:', len(names))\n",
    "                print('\\tnew_features:\\n', names)\n",
    "\n",
    "            new_feature_names.extend(names)\n",
    "            transformer_list.extend([transformer_name] * len(names))\n",
    "        \n",
    "        self.transformer_list, self.column_transformer_features = transformer_list,\\\n",
    "                                                                    new_feature_names\n",
    "\n",
    "        return new_feature_names\n",
    "\n",
    "    \n",
    "    def get_selected_features(self, verbose=None):\n",
    "        \"\"\"\n",
    "\n",
    "        Get the Feature Names that were retained after Feature Selection (sklearn.feature_selection)\n",
    "\n",
    "        Parameters\n",
    "        ----------\n",
    "        verbose : a boolean indicating whether to print summaries. default = False\n",
    "\n",
    "        Returns\n",
    "        -------\n",
    "        a list of the selected feature names\n",
    "\n",
    "\n",
    "        \"\"\"\n",
    "\n",
    "        if verbose is None:\n",
    "            verbose = self.verbose\n",
    "\n",
    "        assert isinstance(self.pipeline, Pipeline), \"Input isn't a Pipeline\"\n",
    "\n",
    "        features = self.get_feature_names()\n",
    "        \n",
    "        if verbose: print('\\n\\n---------\\nRunning get_selected_features\\n---------\\n')\n",
    "            \n",
    "        all_discarded_features, discarding_selectors = [], []\n",
    "\n",
    "        for i, step_item in enumerate(self.pipeline.steps[:]):\n",
    "            \n",
    "            step_name, step = step_item\n",
    "\n",
    "            if hasattr(step, 'get_support'):\n",
    "\n",
    "                if verbose: print('\\nStep ', i, \": \", step_name, ',', \n",
    "                                  step.__class__.__name__, '\\n')\n",
    "                    \n",
    "                check_is_fitted(step)\n",
    "\n",
    "                feature_mask_dict = dict(zip(features, step.get_support()))\n",
    "                \n",
    "                features = [feature for feature, is_retained in feature_mask_dict.items()\\\n",
    "                            if is_retained]\n",
    "                                         \n",
    "                discarded_features = [feature for feature, is_retained in feature_mask_dict.items()\\\n",
    "                                      if not is_retained]\n",
    "                \n",
    "                all_discarded_features.extend(discarded_features)\n",
    "                discarding_selectors.extend([step_name] * len(discarded_features))\n",
    "                \n",
    "                \n",
    "                if verbose: \n",
    "                    print(f'\\t{len(features)} retained, {len(discarded_features)} discarded')\n",
    "                    if len(discarded_features) > 0:\n",
    "                        print('\\n\\tdiscarded_features:\\n\\n', discarded_features)\n",
    "\n",
    "        self.discarded_features, self.discarding_selectors = all_discarded_features,\\\n",
    "                                                                discarding_selectors\n",
    "        \n",
    "        return features\n",
    "\n",
    "    def get_feature_importance(self):\n",
    "        \n",
    "        \"\"\"\n",
    "        Creates a Pandas Series where values are the feature importance values from the model and feature names are set as the index. \n",
    "        \n",
    "        This Series is stored in the `feature_importance` attribute.\n",
    "\n",
    "        Returns\n",
    "        -------\n",
    "        A pandas Series containing the feature importance values and feature names as the index.\n",
    "        \n",
    "        \"\"\"\n",
    "        \n",
    "        assert isinstance(self.pipeline, Pipeline), \"Input isn't a Pipeline\"\n",
    "\n",
    "        features = self.get_selected_features()\n",
    "             \n",
    "        assert hasattr(self.pipeline[-1], 'feature_importances_'),\\\n",
    "            \"The last element in the pipeline isn't an estimator with a feature_importances_ attribute\"\n",
    "        \n",
    "        importance_values = self.pipeline[-1].feature_importances_\n",
    "        \n",
    "        assert len(features) == len(importance_values),\\\n",
    "            \"The number of feature names & importance values doesn't match\"\n",
    "        \n",
    "        feature_importance = pd.Series(importance_values, index=features)\n",
    "        self.feature_importance = feature_importance\n",
    "        \n",
    "        # create feature_info_df\n",
    "        column_transformer_df =\\\n",
    "            pd.DataFrame(dict(transformer=self.transformer_list),\n",
    "                         index=self.column_transformer_features)\n",
    "\n",
    "        discarded_features_df =\\\n",
    "            pd.DataFrame(dict(discarding_selector=self.discarding_selectors),\n",
    "                         index=self.discarded_features)\n",
    "\n",
    "        importance_df = self.feature_importance.rename('value').to_frame()\n",
    "\n",
    "        self.feature_info_df = \\\n",
    "            column_transformer_df\\\n",
    "            .join([importance_df, discarded_features_df])\\\n",
    "            .assign(is_retained = lambda df: ~df.value.isna())        \n",
    "\n",
    "\n",
    "        return feature_importance\n",
    "        \n",
    "    \n",
    "    def plot(self, top_n_features=100, rank_features=True, max_scale=True, \n",
    "             display_imp_values=True, display_imp_value_decimals=1,\n",
    "             height_per_feature=25, orientation='h', width=750, height=None, \n",
    "             str_pad_width=15, yaxes_tickfont_family='Courier New', \n",
    "             yaxes_tickfont_size=15):\n",
    "        \"\"\"\n",
    "\n",
    "        Plot the Feature Names & Importances \n",
    "\n",
    "\n",
    "        Parameters\n",
    "        ----------\n",
    "\n",
    "        top_n_features : the number of features to plot, default is 100\n",
    "        rank_features : whether to rank the features with integers, default is True\n",
    "        max_scale : Should the importance values be scaled by the maximum value & mulitplied by 100?  Default is True.\n",
    "        display_imp_values : Should the importance values be displayed? Default is True.\n",
    "        display_imp_value_decimals : If display_imp_values is True, how many decimal places should be displayed. Default is 1.\n",
    "        height_per_feature : if height is None, the plot height is calculated by top_n_features * height_per_feature. \n",
    "        This allows all the features enough space to be displayed\n",
    "        orientation : the plot orientation, 'h' (default) or 'v'\n",
    "        width :  the width of the plot, default is 500\n",
    "        height : the height of the plot, the default is top_n_features * height_per_feature\n",
    "        str_pad_width : When rank_features=True, this number of spaces to add between the rank integer and feature name. \n",
    "            This will enable the rank integers to line up with each other for easier reading. \n",
    "            Default is 15. If you have long feature names, you can increase this number to make the integers line up more.\n",
    "            It can also be set to 0.\n",
    "        yaxes_tickfont_family : the font for the feature names. Default is Courier New.\n",
    "        yaxes_tickfont_size : the font size for the feature names. Default is 15.\n",
    "\n",
    "        Returns\n",
    "        -------\n",
    "        plot\n",
    "\n",
    "        \"\"\"\n",
    "        if height is None:\n",
    "            height = top_n_features * height_per_feature\n",
    "            \n",
    "        # prep the data\n",
    "        \n",
    "        all_importances = self.get_feature_importance()\n",
    "        n_all_importances = len(all_importances)\n",
    "        \n",
    "        plot_importances_df =\\\n",
    "            all_importances\\\n",
    "            .nlargest(top_n_features)\\\n",
    "            .sort_values()\\\n",
    "            .to_frame('value')\\\n",
    "            .rename_axis('feature')\\\n",
    "            .reset_index()\n",
    "                \n",
    "        if max_scale:\n",
    "            plot_importances_df['value'] = \\\n",
    "                                plot_importances_df.value.abs() /\\\n",
    "                                plot_importances_df.value.abs().max() * 100\n",
    "            \n",
    "        self.plot_importances_df = plot_importances_df.copy()\n",
    "        \n",
    "        if len(all_importances) < top_n_features:\n",
    "            title_text = 'All Feature Importances'\n",
    "        else:\n",
    "            title_text = f'Top {top_n_features} (of {n_all_importances}) Feature Importances'       \n",
    "        \n",
    "        if rank_features:\n",
    "            padded_features = \\\n",
    "                plot_importances_df.feature\\\n",
    "                .str.pad(width=str_pad_width)\\\n",
    "                .values\n",
    "            \n",
    "            ranked_features =\\\n",
    "                plot_importances_df.index\\\n",
    "                .to_series()\\\n",
    "                .sort_values(ascending=False)\\\n",
    "                .add(1)\\\n",
    "                .astype(str)\\\n",
    "                .str.cat(padded_features, sep='. ')\\\n",
    "                .values\n",
    "\n",
    "            plot_importances_df['feature'] = ranked_features\n",
    "        \n",
    "        if display_imp_values:\n",
    "            text = plot_importances_df.value.round(display_imp_value_decimals)\n",
    "        else:\n",
    "            text = None\n",
    "\n",
    "        # create the plot \n",
    "        \n",
    "        fig = px.bar(plot_importances_df, \n",
    "                     x='value', \n",
    "                     y='feature',\n",
    "                     orientation=orientation, \n",
    "                     width=width, \n",
    "                     height=height,\n",
    "                     text=text)\n",
    "        fig.update_layout(title_text=title_text, title_x=0.5) \n",
    "        fig.update(layout_showlegend=False)\n",
    "        fig.update_yaxes(tickfont=dict(family=yaxes_tickfont_family, \n",
    "                                       size=yaxes_tickfont_size),\n",
    "                         title='')\n",
    "        fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56b650ac",
   "metadata": {},
   "source": [
    "### Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 746,
   "id": "091d213d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/proctor/.pyenv/versions/3.8.12/envs/lewagon/lib/python3.8/site-packages/sklearn/utils/deprecation.py:87: FutureWarning:\n",
      "\n",
      "Function get_feature_names is deprecated; get_feature_names is deprecated in 1.0 and will be removed in 1.2. Please use get_feature_names_out instead.\n",
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "plotlyServerURL": "https://plot.ly"
       },
       "data": [
        {
         "alignmentgroup": "True",
         "hovertemplate": "value=%{x}<br>feature=%{y}<br>text=%{text}<extra></extra>",
         "legendgroup": "",
         "marker": {
          "color": "#636efa",
          "pattern": {
           "shape": ""
          }
         },
         "name": "",
         "offsetgroup": "",
         "orientation": "h",
         "showlegend": false,
         "text": [
          0.3,
          0.5,
          1.3,
          2.6,
          2.7,
          29.7,
          42.1,
          42.4,
          50,
          100
         ],
         "textposition": "auto",
         "type": "bar",
         "x": [
          0.33325229380316784,
          0.5070120187367956,
          1.2779936661944686,
          2.5510217917430373,
          2.7496563027460987,
          29.65926321120671,
          42.08971175203908,
          42.42973944366756,
          50.01937447502013,
          100
         ],
         "xaxis": "x",
         "y": [
          "10.       x0_Casado",
          "9.         vigente",
          "8.      x0_Soltero",
          "7.           x1_Si",
          "6.           x1_No",
          "5. nro_cuotas_atrasadas",
          "4.           monto",
          "3. cuotas_vencidas",
          "2. cuotas_canceladas",
          "1.     dias_atraso"
         ],
         "yaxis": "y"
        }
       ],
       "layout": {
        "barmode": "relative",
        "height": 625,
        "legend": {
         "tracegroupgap": 0
        },
        "margin": {
         "t": 60
        },
        "showlegend": false,
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "pie": [
           {
            "automargin": true,
            "type": "pie"
           }
          ],
          "scatter": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "autotypenumbers": "strict",
          "coloraxis": {
           "colorbar": {
            "outlinewidth": 0,
            "ticks": ""
           }
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        },
        "title": {
         "text": "All Feature Importances",
         "x": 0.5
        },
        "width": 750,
        "xaxis": {
         "anchor": "y",
         "domain": [
          0,
          1
         ],
         "title": {
          "text": "value"
         }
        },
        "yaxis": {
         "anchor": "x",
         "domain": [
          0,
          1
         ],
         "tickfont": {
          "family": "Courier New",
          "size": 15
         },
         "title": {
          "text": ""
         }
        }
       }
      },
      "text/html": [
       "<div>                            <div id=\"466e0d47-3642-4cf0-9692-ee66f10ba791\" class=\"plotly-graph-div\" style=\"height:625px; width:750px;\"></div>            <script type=\"text/javascript\">                require([\"plotly\"], function(Plotly) {                    window.PLOTLYENV=window.PLOTLYENV || {};                                    if (document.getElementById(\"466e0d47-3642-4cf0-9692-ee66f10ba791\")) {                    Plotly.newPlot(                        \"466e0d47-3642-4cf0-9692-ee66f10ba791\",                        [{\"alignmentgroup\":\"True\",\"hovertemplate\":\"value=%{x}<br>feature=%{y}<br>text=%{text}<extra></extra>\",\"legendgroup\":\"\",\"marker\":{\"color\":\"#636efa\",\"pattern\":{\"shape\":\"\"}},\"name\":\"\",\"offsetgroup\":\"\",\"orientation\":\"h\",\"showlegend\":false,\"text\":[0.3,0.5,1.3,2.6,2.7,29.7,42.1,42.4,50.0,100.0],\"textposition\":\"auto\",\"x\":[0.33325229380316784,0.5070120187367956,1.2779936661944686,2.5510217917430373,2.7496563027460987,29.65926321120671,42.08971175203908,42.42973944366756,50.01937447502013,100.0],\"xaxis\":\"x\",\"y\":[\"10.       x0_Casado\",\"9.         vigente\",\"8.      x0_Soltero\",\"7.           x1_Si\",\"6.           x1_No\",\"5. nro_cuotas_atrasadas\",\"4.           monto\",\"3. cuotas_vencidas\",\"2. cuotas_canceladas\",\"1.     dias_atraso\"],\"yaxis\":\"y\",\"type\":\"bar\"}],                        {\"template\":{\"data\":{\"bar\":[{\"error_x\":{\"color\":\"#2a3f5f\"},\"error_y\":{\"color\":\"#2a3f5f\"},\"marker\":{\"line\":{\"color\":\"#E5ECF6\",\"width\":0.5},\"pattern\":{\"fillmode\":\"overlay\",\"size\":10,\"solidity\":0.2}},\"type\":\"bar\"}],\"barpolar\":[{\"marker\":{\"line\":{\"color\":\"#E5ECF6\",\"width\":0.5},\"pattern\":{\"fillmode\":\"overlay\",\"size\":10,\"solidity\":0.2}},\"type\":\"barpolar\"}],\"carpet\":[{\"aaxis\":{\"endlinecolor\":\"#2a3f5f\",\"gridcolor\":\"white\",\"linecolor\":\"white\",\"minorgridcolor\":\"white\",\"startlinecolor\":\"#2a3f5f\"},\"baxis\":{\"endlinecolor\":\"#2a3f5f\",\"gridcolor\":\"white\",\"linecolor\":\"white\",\"minorgridcolor\":\"white\",\"startlinecolor\":\"#2a3f5f\"},\"type\":\"carpet\"}],\"choropleth\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"type\":\"choropleth\"}],\"contour\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"colorscale\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]],\"type\":\"contour\"}],\"contourcarpet\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"type\":\"contourcarpet\"}],\"heatmap\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"colorscale\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]],\"type\":\"heatmap\"}],\"heatmapgl\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"colorscale\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]],\"type\":\"heatmapgl\"}],\"histogram\":[{\"marker\":{\"pattern\":{\"fillmode\":\"overlay\",\"size\":10,\"solidity\":0.2}},\"type\":\"histogram\"}],\"histogram2d\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"colorscale\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]],\"type\":\"histogram2d\"}],\"histogram2dcontour\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"colorscale\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]],\"type\":\"histogram2dcontour\"}],\"mesh3d\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"type\":\"mesh3d\"}],\"parcoords\":[{\"line\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"parcoords\"}],\"pie\":[{\"automargin\":true,\"type\":\"pie\"}],\"scatter\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scatter\"}],\"scatter3d\":[{\"line\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scatter3d\"}],\"scattercarpet\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scattercarpet\"}],\"scattergeo\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scattergeo\"}],\"scattergl\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scattergl\"}],\"scattermapbox\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scattermapbox\"}],\"scatterpolar\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scatterpolar\"}],\"scatterpolargl\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scatterpolargl\"}],\"scatterternary\":[{\"marker\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"type\":\"scatterternary\"}],\"surface\":[{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"},\"colorscale\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]],\"type\":\"surface\"}],\"table\":[{\"cells\":{\"fill\":{\"color\":\"#EBF0F8\"},\"line\":{\"color\":\"white\"}},\"header\":{\"fill\":{\"color\":\"#C8D4E3\"},\"line\":{\"color\":\"white\"}},\"type\":\"table\"}]},\"layout\":{\"annotationdefaults\":{\"arrowcolor\":\"#2a3f5f\",\"arrowhead\":0,\"arrowwidth\":1},\"autotypenumbers\":\"strict\",\"coloraxis\":{\"colorbar\":{\"outlinewidth\":0,\"ticks\":\"\"}},\"colorscale\":{\"diverging\":[[0,\"#8e0152\"],[0.1,\"#c51b7d\"],[0.2,\"#de77ae\"],[0.3,\"#f1b6da\"],[0.4,\"#fde0ef\"],[0.5,\"#f7f7f7\"],[0.6,\"#e6f5d0\"],[0.7,\"#b8e186\"],[0.8,\"#7fbc41\"],[0.9,\"#4d9221\"],[1,\"#276419\"]],\"sequential\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]],\"sequentialminus\":[[0.0,\"#0d0887\"],[0.1111111111111111,\"#46039f\"],[0.2222222222222222,\"#7201a8\"],[0.3333333333333333,\"#9c179e\"],[0.4444444444444444,\"#bd3786\"],[0.5555555555555556,\"#d8576b\"],[0.6666666666666666,\"#ed7953\"],[0.7777777777777778,\"#fb9f3a\"],[0.8888888888888888,\"#fdca26\"],[1.0,\"#f0f921\"]]},\"colorway\":[\"#636efa\",\"#EF553B\",\"#00cc96\",\"#ab63fa\",\"#FFA15A\",\"#19d3f3\",\"#FF6692\",\"#B6E880\",\"#FF97FF\",\"#FECB52\"],\"font\":{\"color\":\"#2a3f5f\"},\"geo\":{\"bgcolor\":\"white\",\"lakecolor\":\"white\",\"landcolor\":\"#E5ECF6\",\"showlakes\":true,\"showland\":true,\"subunitcolor\":\"white\"},\"hoverlabel\":{\"align\":\"left\"},\"hovermode\":\"closest\",\"mapbox\":{\"style\":\"light\"},\"paper_bgcolor\":\"white\",\"plot_bgcolor\":\"#E5ECF6\",\"polar\":{\"angularaxis\":{\"gridcolor\":\"white\",\"linecolor\":\"white\",\"ticks\":\"\"},\"bgcolor\":\"#E5ECF6\",\"radialaxis\":{\"gridcolor\":\"white\",\"linecolor\":\"white\",\"ticks\":\"\"}},\"scene\":{\"xaxis\":{\"backgroundcolor\":\"#E5ECF6\",\"gridcolor\":\"white\",\"gridwidth\":2,\"linecolor\":\"white\",\"showbackground\":true,\"ticks\":\"\",\"zerolinecolor\":\"white\"},\"yaxis\":{\"backgroundcolor\":\"#E5ECF6\",\"gridcolor\":\"white\",\"gridwidth\":2,\"linecolor\":\"white\",\"showbackground\":true,\"ticks\":\"\",\"zerolinecolor\":\"white\"},\"zaxis\":{\"backgroundcolor\":\"#E5ECF6\",\"gridcolor\":\"white\",\"gridwidth\":2,\"linecolor\":\"white\",\"showbackground\":true,\"ticks\":\"\",\"zerolinecolor\":\"white\"}},\"shapedefaults\":{\"line\":{\"color\":\"#2a3f5f\"}},\"ternary\":{\"aaxis\":{\"gridcolor\":\"white\",\"linecolor\":\"white\",\"ticks\":\"\"},\"baxis\":{\"gridcolor\":\"white\",\"linecolor\":\"white\",\"ticks\":\"\"},\"bgcolor\":\"#E5ECF6\",\"caxis\":{\"gridcolor\":\"white\",\"linecolor\":\"white\",\"ticks\":\"\"}},\"title\":{\"x\":0.05},\"xaxis\":{\"automargin\":true,\"gridcolor\":\"white\",\"linecolor\":\"white\",\"ticks\":\"\",\"title\":{\"standoff\":15},\"zerolinecolor\":\"white\",\"zerolinewidth\":2},\"yaxis\":{\"automargin\":true,\"gridcolor\":\"white\",\"linecolor\":\"white\",\"ticks\":\"\",\"title\":{\"standoff\":15},\"zerolinecolor\":\"white\",\"zerolinewidth\":2}}},\"xaxis\":{\"anchor\":\"y\",\"domain\":[0.0,1.0],\"title\":{\"text\":\"value\"}},\"yaxis\":{\"anchor\":\"x\",\"domain\":[0.0,1.0],\"title\":{\"text\":\"\"},\"tickfont\":{\"family\":\"Courier New\",\"size\":15}},\"legend\":{\"tracegroupgap\":0},\"margin\":{\"t\":60},\"barmode\":\"relative\",\"height\":625,\"width\":750,\"title\":{\"text\":\"All Feature Importances\",\"x\":0.5},\"showlegend\":false},                        {\"responsive\": true}                    ).then(function(){\n",
       "                            \n",
       "var gd = document.getElementById('466e0d47-3642-4cf0-9692-ee66f10ba791');\n",
       "var x = new MutationObserver(function (mutations, observer) {{\n",
       "        var display = window.getComputedStyle(gd).display;\n",
       "        if (!display || display === 'none') {{\n",
       "            console.log([gd, 'removed!']);\n",
       "            Plotly.purge(gd);\n",
       "            observer.disconnect();\n",
       "        }}\n",
       "}});\n",
       "\n",
       "// Listen for the removal of the full notebook cells\n",
       "var notebookContainer = gd.closest('#notebook-container');\n",
       "if (notebookContainer) {{\n",
       "    x.observe(notebookContainer, {childList: true});\n",
       "}}\n",
       "\n",
       "// Listen for the clearing of the current output cell\n",
       "var outputEl = gd.closest('.output');\n",
       "if (outputEl) {{\n",
       "    x.observe(outputEl, {childList: true});\n",
       "}}\n",
       "\n",
       "                        })                };                });            </script>        </div>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "feature_importance = FeatureImportance(pipe_baseline)\n",
    "feature_importance.plot(top_n_features=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3297892b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f8c692e0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c6688e5e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1d8f2033",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58f06217",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a622d7b6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8392fff7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c88193c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "9dab0f83",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "###### Code Tests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d3792328",
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "model = GradientBoostingRegressor()\n",
    "cross_validate(model, X_train, y_test, cv = 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 555,
   "id": "26d455a1",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.12145103 0.08340243 0.05080577 0.40709578 0.30602226 0.0125019\n",
      " 0.00150969 0.00052677 0.00534452 0.01133986]\n"
     ]
    }
   ],
   "source": [
    "model = GradientBoostingRegressor()\n",
    "model.fit(X_transformed, y_train)\n",
    "# feature importance\n",
    "print(model.feature_importances_)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 569,
   "id": "d41c155e",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.067113</td>\n",
       "      <td>0.197405</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.190476</td>\n",
       "      <td>0.270833</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.042149</td>\n",
       "      <td>0.056103</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>0.125000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.116860</td>\n",
       "      <td>0.028997</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.095238</td>\n",
       "      <td>0.041667</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.058792</td>\n",
       "      <td>0.593508</td>\n",
       "      <td>0.75</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>0.229167</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.477026</td>\n",
       "      <td>0.072422</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.071429</td>\n",
       "      <td>0.437500</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10189</th>\n",
       "      <td>0.024602</td>\n",
       "      <td>0.070233</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.071429</td>\n",
       "      <td>0.104167</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10190</th>\n",
       "      <td>0.033466</td>\n",
       "      <td>0.466438</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.145833</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10191</th>\n",
       "      <td>0.095695</td>\n",
       "      <td>0.084363</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.476190</td>\n",
       "      <td>0.083333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10192</th>\n",
       "      <td>0.067836</td>\n",
       "      <td>0.005800</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.023810</td>\n",
       "      <td>0.270833</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10193</th>\n",
       "      <td>0.034913</td>\n",
       "      <td>0.087189</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>0.145833</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10194 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              0         1     2         3         4    5    6    7    8    9\n",
       "0      0.067113  0.197405  0.40  0.190476  0.270833  1.0  1.0  0.0  1.0  0.0\n",
       "1      0.042149  0.056103  0.40  0.428571  0.125000  0.0  0.0  1.0  0.0  1.0\n",
       "2      0.116860  0.028997  0.10  0.095238  0.041667  0.0  1.0  0.0  1.0  0.0\n",
       "3      0.058792  0.593508  0.75  0.428571  0.229167  1.0  1.0  0.0  0.0  1.0\n",
       "4      0.477026  0.072422  0.40  0.071429  0.437500  0.0  1.0  0.0  0.0  1.0\n",
       "...         ...       ...   ...       ...       ...  ...  ...  ...  ...  ...\n",
       "10189  0.024602  0.070233  0.40  0.071429  0.104167  0.0  0.0  1.0  1.0  0.0\n",
       "10190  0.033466  0.466438  0.40  0.000000  0.145833  0.0  0.0  1.0  1.0  0.0\n",
       "10191  0.095695  0.084363  0.40  0.476190  0.083333  0.0  1.0  0.0  1.0  0.0\n",
       "10192  0.067836  0.005800  0.00  0.023810  0.270833  0.0  1.0  0.0  1.0  0.0\n",
       "10193  0.034913  0.087189  0.40  0.142857  0.145833  0.0  0.0  1.0  0.0  1.0\n",
       "\n",
       "[10194 rows x 10 columns]"
      ]
     },
     "execution_count": 569,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_transformed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 576,
   "id": "f5557303",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "# plot\n",
    "plt.bar(range(len(model.feature_importances_)), model.feature_importances_)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 567,
   "id": "6aabef0c",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.12145103 0.08340243 0.05080577 0.40709578 0.30602226 0.0125019\n",
      " 0.00150969 0.00052677 0.00534452 0.01133986]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(model.feature_importances_) #use inbuilt class feature_importances of tree based classifiers\n",
    "#plot graph of feature importances for better visualization\n",
    "feat_importances = pd.Series(model.feature_importances_, index=X_transformed.columns)\n",
    "feat_importances.nlargest(10).plot(kind='barh')\n",
    "plt.show()\n",
    "plt.yticks(pos, np.array(diabetes.feature_names)[sorted_idx])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 577,
   "id": "51ae3d18",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "tree must be Booster, XGBModel or dict instance",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m/tmp/ipykernel_769/932148123.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mxgboost\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mplot_importance\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mplot_importance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m \u001b[0mpyplot\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.pyenv/versions/3.8.12/envs/lewagon/lib/python3.8/site-packages/xgboost/plotting.py\u001b[0m in \u001b[0;36mplot_importance\u001b[0;34m(booster, ax, height, xlim, ylim, title, xlabel, ylabel, fmap, importance_type, max_num_features, grid, show_values, **kwargs)\u001b[0m\n\u001b[1;32m     69\u001b[0m         \u001b[0mimportance\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbooster\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     70\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 71\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'tree must be Booster, XGBModel or dict instance'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     72\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     73\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mimportance\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: tree must be Booster, XGBModel or dict instance"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from xgboost import plot_importance\n",
    "\n",
    "plot_importance(model)\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 581,
   "id": "891a4c1a",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'GradientBoostingRegressor' object has no attribute 'feature_names'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m/tmp/ipykernel_769/3332551834.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msubplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbarh\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpos\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeature_importance\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msorted_idx\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malign\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"center\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0myticks\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpos\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfeature_names\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msorted_idx\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      8\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Feature Importance (MDI)\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'GradientBoostingRegressor' object has no attribute 'feature_names'"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_importance = model.feature_importances_\n",
    "sorted_idx = np.argsort(feature_importance)\n",
    "pos = np.arange(sorted_idx.shape[0]) + 0.5\n",
    "fig = plt.figure(figsize=(12, 6))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.barh(pos, feature_importance[sorted_idx], align=\"center\")\n",
    "plt.yticks(pos, np.array(model.feature_names)[sorted_idx])\n",
    "plt.title(\"Feature Importance (MDI)\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.12"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "409.6px"
   },
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}