salary_pred.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <span style=\"color:gray\">Salary Prediction using Linear Regression & Gradient Boosting Regressor</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## <span style=\"color:black\">1. Define</span>   \n",
    " Problem -  \n",
    " >Predict Salary  \n",
    " \n",
    " What we have - \n",
    " >100k observations.  \n",
    " >7 features - companyId, jobType, degree, major, industry, yearsExperience, milesFromMetropolis  \n",
    " >1 target - salary  \n",
    " >4 Categorical/Qualitative - companyId, jobType, degree, major, industry  \n",
    " >2 Numeric/Quantitative - yearsExperience, milesFromMetropolis  \n",
    " \n",
    " Type of Algorithm -  \n",
    " >Linear Regression  \n",
    " >Gradient Boosting Regressor"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## <span style=\"color:black\">2. Discover</span> "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">2.1. Import Libraries</span> "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture\n",
    "! pip install seaborn\n",
    "#import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline \n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">2.2. Import data</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_features_df = pd.read_csv(\"data/train_features.csv\")\n",
    "train_target_df = pd.read_csv(\"data/train_salaries.csv\")\n",
    "test_features_df = pd.read_csv(\"data/test_features.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">2.3. Clean data</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>jobId</th>\n",
       "      <th>companyId</th>\n",
       "      <th>jobType</th>\n",
       "      <th>degree</th>\n",
       "      <th>major</th>\n",
       "      <th>industry</th>\n",
       "      <th>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>JOB1362684407687</td>\n",
       "      <td>COMP37</td>\n",
       "      <td>CFO</td>\n",
       "      <td>MASTERS</td>\n",
       "      <td>MATH</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>10</td>\n",
       "      <td>83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>JOB1362684407688</td>\n",
       "      <td>COMP19</td>\n",
       "      <td>CEO</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>WEB</td>\n",
       "      <td>3</td>\n",
       "      <td>73</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>JOB1362684407689</td>\n",
       "      <td>COMP52</td>\n",
       "      <td>VICE_PRESIDENT</td>\n",
       "      <td>DOCTORAL</td>\n",
       "      <td>PHYSICS</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>10</td>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>JOB1362684407690</td>\n",
       "      <td>COMP38</td>\n",
       "      <td>MANAGER</td>\n",
       "      <td>DOCTORAL</td>\n",
       "      <td>CHEMISTRY</td>\n",
       "      <td>AUTO</td>\n",
       "      <td>8</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>JOB1362684407691</td>\n",
       "      <td>COMP7</td>\n",
       "      <td>VICE_PRESIDENT</td>\n",
       "      <td>BACHELORS</td>\n",
       "      <td>PHYSICS</td>\n",
       "      <td>FINANCE</td>\n",
       "      <td>8</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              jobId companyId         jobType       degree      major  \\\n",
       "0  JOB1362684407687    COMP37             CFO      MASTERS       MATH   \n",
       "1  JOB1362684407688    COMP19             CEO  HIGH_SCHOOL       NONE   \n",
       "2  JOB1362684407689    COMP52  VICE_PRESIDENT     DOCTORAL    PHYSICS   \n",
       "3  JOB1362684407690    COMP38         MANAGER     DOCTORAL  CHEMISTRY   \n",
       "4  JOB1362684407691     COMP7  VICE_PRESIDENT    BACHELORS    PHYSICS   \n",
       "\n",
       "  industry  yearsExperience  milesFromMetropolis  \n",
       "0   HEALTH               10                   83  \n",
       "1      WEB                3                   73  \n",
       "2   HEALTH               10                   38  \n",
       "3     AUTO                8                   17  \n",
       "4  FINANCE                8                   16  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_features_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>jobId</th>\n",
       "      <th>salary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>JOB1362684407687</td>\n",
       "      <td>130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>JOB1362684407688</td>\n",
       "      <td>101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>JOB1362684407689</td>\n",
       "      <td>137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>JOB1362684407690</td>\n",
       "      <td>142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>JOB1362684407691</td>\n",
       "      <td>163</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              jobId  salary\n",
       "0  JOB1362684407687     130\n",
       "1  JOB1362684407688     101\n",
       "2  JOB1362684407689     137\n",
       "3  JOB1362684407690     142\n",
       "4  JOB1362684407691     163"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_target_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>jobId</th>\n",
       "      <th>companyId</th>\n",
       "      <th>jobType</th>\n",
       "      <th>degree</th>\n",
       "      <th>major</th>\n",
       "      <th>industry</th>\n",
       "      <th>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>JOB1362685407687</td>\n",
       "      <td>COMP33</td>\n",
       "      <td>MANAGER</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>22</td>\n",
       "      <td>73</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>JOB1362685407688</td>\n",
       "      <td>COMP13</td>\n",
       "      <td>JUNIOR</td>\n",
       "      <td>NONE</td>\n",
       "      <td>NONE</td>\n",
       "      <td>AUTO</td>\n",
       "      <td>20</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>JOB1362685407689</td>\n",
       "      <td>COMP10</td>\n",
       "      <td>CTO</td>\n",
       "      <td>MASTERS</td>\n",
       "      <td>BIOLOGY</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>17</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>JOB1362685407690</td>\n",
       "      <td>COMP21</td>\n",
       "      <td>MANAGER</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>OIL</td>\n",
       "      <td>14</td>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>JOB1362685407691</td>\n",
       "      <td>COMP36</td>\n",
       "      <td>JUNIOR</td>\n",
       "      <td>DOCTORAL</td>\n",
       "      <td>BIOLOGY</td>\n",
       "      <td>OIL</td>\n",
       "      <td>10</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              jobId companyId  jobType       degree    major industry  \\\n",
       "0  JOB1362685407687    COMP33  MANAGER  HIGH_SCHOOL     NONE   HEALTH   \n",
       "1  JOB1362685407688    COMP13   JUNIOR         NONE     NONE     AUTO   \n",
       "2  JOB1362685407689    COMP10      CTO      MASTERS  BIOLOGY   HEALTH   \n",
       "3  JOB1362685407690    COMP21  MANAGER  HIGH_SCHOOL     NONE      OIL   \n",
       "4  JOB1362685407691    COMP36   JUNIOR     DOCTORAL  BIOLOGY      OIL   \n",
       "\n",
       "   yearsExperience  milesFromMetropolis  \n",
       "0               22                   73  \n",
       "1               20                   47  \n",
       "2               17                    9  \n",
       "3               14                   96  \n",
       "4               10                   44  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_features_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#join the train set with target on jobId as it looks like its a common set\n",
    "train_merge = pd.merge(train_features_df, train_target_df, on='jobId')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>jobId</th>\n",
       "      <th>companyId</th>\n",
       "      <th>jobType</th>\n",
       "      <th>degree</th>\n",
       "      <th>major</th>\n",
       "      <th>industry</th>\n",
       "      <th>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "      <th>salary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>JOB1362684407687</td>\n",
       "      <td>COMP37</td>\n",
       "      <td>CFO</td>\n",
       "      <td>MASTERS</td>\n",
       "      <td>MATH</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>10</td>\n",
       "      <td>83</td>\n",
       "      <td>130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>JOB1362684407688</td>\n",
       "      <td>COMP19</td>\n",
       "      <td>CEO</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>WEB</td>\n",
       "      <td>3</td>\n",
       "      <td>73</td>\n",
       "      <td>101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>JOB1362684407689</td>\n",
       "      <td>COMP52</td>\n",
       "      <td>VICE_PRESIDENT</td>\n",
       "      <td>DOCTORAL</td>\n",
       "      <td>PHYSICS</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>10</td>\n",
       "      <td>38</td>\n",
       "      <td>137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>JOB1362684407690</td>\n",
       "      <td>COMP38</td>\n",
       "      <td>MANAGER</td>\n",
       "      <td>DOCTORAL</td>\n",
       "      <td>CHEMISTRY</td>\n",
       "      <td>AUTO</td>\n",
       "      <td>8</td>\n",
       "      <td>17</td>\n",
       "      <td>142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>JOB1362684407691</td>\n",
       "      <td>COMP7</td>\n",
       "      <td>VICE_PRESIDENT</td>\n",
       "      <td>BACHELORS</td>\n",
       "      <td>PHYSICS</td>\n",
       "      <td>FINANCE</td>\n",
       "      <td>8</td>\n",
       "      <td>16</td>\n",
       "      <td>163</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              jobId companyId         jobType       degree      major  \\\n",
       "0  JOB1362684407687    COMP37             CFO      MASTERS       MATH   \n",
       "1  JOB1362684407688    COMP19             CEO  HIGH_SCHOOL       NONE   \n",
       "2  JOB1362684407689    COMP52  VICE_PRESIDENT     DOCTORAL    PHYSICS   \n",
       "3  JOB1362684407690    COMP38         MANAGER     DOCTORAL  CHEMISTRY   \n",
       "4  JOB1362684407691     COMP7  VICE_PRESIDENT    BACHELORS    PHYSICS   \n",
       "\n",
       "  industry  yearsExperience  milesFromMetropolis  salary  \n",
       "0   HEALTH               10                   83     130  \n",
       "1      WEB                3                   73     101  \n",
       "2   HEALTH               10                   38     137  \n",
       "3     AUTO                8                   17     142  \n",
       "4  FINANCE                8                   16     163  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_merge.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 1000000 entries, 0 to 999999\n",
      "Data columns (total 9 columns):\n",
      " #   Column               Non-Null Count    Dtype \n",
      "---  ------               --------------    ----- \n",
      " 0   jobId                1000000 non-null  object\n",
      " 1   companyId            1000000 non-null  object\n",
      " 2   jobType              1000000 non-null  object\n",
      " 3   degree               1000000 non-null  object\n",
      " 4   major                1000000 non-null  object\n",
      " 5   industry             1000000 non-null  object\n",
      " 6   yearsExperience      1000000 non-null  int64 \n",
      " 7   milesFromMetropolis  1000000 non-null  int64 \n",
      " 8   salary               1000000 non-null  int64 \n",
      "dtypes: int64(3), object(6)\n",
      "memory usage: 76.3+ MB\n"
     ]
    }
   ],
   "source": [
    "# check datatypes\n",
    "train_merge.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "jobId                  0\n",
       "companyId              0\n",
       "jobType                0\n",
       "degree                 0\n",
       "major                  0\n",
       "industry               0\n",
       "yearsExperience        0\n",
       "milesFromMetropolis    0\n",
       "salary                 0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#check for nulls \n",
    "train_merge.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>jobId</th>\n",
       "      <th>companyId</th>\n",
       "      <th>jobType</th>\n",
       "      <th>degree</th>\n",
       "      <th>major</th>\n",
       "      <th>industry</th>\n",
       "      <th>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "      <th>salary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1000000</td>\n",
       "      <td>1000000</td>\n",
       "      <td>1000000</td>\n",
       "      <td>1000000</td>\n",
       "      <td>1000000</td>\n",
       "      <td>1000000</td>\n",
       "      <td>1000000.000000</td>\n",
       "      <td>1000000.000000</td>\n",
       "      <td>1000000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>1000000</td>\n",
       "      <td>63</td>\n",
       "      <td>8</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>7</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>JOB1362684550716</td>\n",
       "      <td>COMP39</td>\n",
       "      <td>SENIOR</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>WEB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>1</td>\n",
       "      <td>16193</td>\n",
       "      <td>125886</td>\n",
       "      <td>236976</td>\n",
       "      <td>532355</td>\n",
       "      <td>143206</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.992386</td>\n",
       "      <td>49.529260</td>\n",
       "      <td>116.061818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.212391</td>\n",
       "      <td>28.877733</td>\n",
       "      <td>38.717936</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>88.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>114.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>141.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>99.000000</td>\n",
       "      <td>301.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   jobId companyId  jobType       degree    major industry  \\\n",
       "count            1000000   1000000  1000000      1000000  1000000  1000000   \n",
       "unique           1000000        63        8            5        9        7   \n",
       "top     JOB1362684550716    COMP39   SENIOR  HIGH_SCHOOL     NONE      WEB   \n",
       "freq                   1     16193   125886       236976   532355   143206   \n",
       "mean                 NaN       NaN      NaN          NaN      NaN      NaN   \n",
       "std                  NaN       NaN      NaN          NaN      NaN      NaN   \n",
       "min                  NaN       NaN      NaN          NaN      NaN      NaN   \n",
       "25%                  NaN       NaN      NaN          NaN      NaN      NaN   \n",
       "50%                  NaN       NaN      NaN          NaN      NaN      NaN   \n",
       "75%                  NaN       NaN      NaN          NaN      NaN      NaN   \n",
       "max                  NaN       NaN      NaN          NaN      NaN      NaN   \n",
       "\n",
       "        yearsExperience  milesFromMetropolis          salary  \n",
       "count    1000000.000000       1000000.000000  1000000.000000  \n",
       "unique              NaN                  NaN             NaN  \n",
       "top                 NaN                  NaN             NaN  \n",
       "freq                NaN                  NaN             NaN  \n",
       "mean          11.992386            49.529260      116.061818  \n",
       "std            7.212391            28.877733       38.717936  \n",
       "min            0.000000             0.000000        0.000000  \n",
       "25%            6.000000            25.000000       88.000000  \n",
       "50%           12.000000            50.000000      114.000000  \n",
       "75%           18.000000            75.000000      141.000000  \n",
       "max           24.000000            99.000000      301.000000  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_merge.describe(include = 'all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#dropping records with Salaries == 0\n",
    "train_merge = train_merge[train_merge.salary != 0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### <span style=\"color:black\">2.1. Import Libraries</span><Font color = black>  2.3.1 Observations</font>  \n",
    "> dont see any nulls.  \n",
    "> there are some observations that have 0 salaries, which seems an error, so dropped.  \n",
    "> overall data looks clean."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">2.1. Import Libraries</span> "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>jobId</th>\n",
       "      <th>companyId</th>\n",
       "      <th>jobType</th>\n",
       "      <th>degree</th>\n",
       "      <th>major</th>\n",
       "      <th>industry</th>\n",
       "      <th>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "      <th>salary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>999995</td>\n",
       "      <td>999995</td>\n",
       "      <td>999995</td>\n",
       "      <td>999995</td>\n",
       "      <td>999995</td>\n",
       "      <td>999995</td>\n",
       "      <td>999995.000000</td>\n",
       "      <td>999995.000000</td>\n",
       "      <td>999995.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>999995</td>\n",
       "      <td>63</td>\n",
       "      <td>8</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>7</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>JOB1362684550716</td>\n",
       "      <td>COMP39</td>\n",
       "      <td>SENIOR</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>WEB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>1</td>\n",
       "      <td>16193</td>\n",
       "      <td>125886</td>\n",
       "      <td>236975</td>\n",
       "      <td>532353</td>\n",
       "      <td>143205</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.992407</td>\n",
       "      <td>49.529381</td>\n",
       "      <td>116.062398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.212390</td>\n",
       "      <td>28.877721</td>\n",
       "      <td>38.717163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>17.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>88.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>114.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>141.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>99.000000</td>\n",
       "      <td>301.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   jobId companyId jobType       degree   major industry  \\\n",
       "count             999995    999995  999995       999995  999995   999995   \n",
       "unique            999995        63       8            5       9        7   \n",
       "top     JOB1362684550716    COMP39  SENIOR  HIGH_SCHOOL    NONE      WEB   \n",
       "freq                   1     16193  125886       236975  532353   143205   \n",
       "mean                 NaN       NaN     NaN          NaN     NaN      NaN   \n",
       "std                  NaN       NaN     NaN          NaN     NaN      NaN   \n",
       "min                  NaN       NaN     NaN          NaN     NaN      NaN   \n",
       "25%                  NaN       NaN     NaN          NaN     NaN      NaN   \n",
       "50%                  NaN       NaN     NaN          NaN     NaN      NaN   \n",
       "75%                  NaN       NaN     NaN          NaN     NaN      NaN   \n",
       "max                  NaN       NaN     NaN          NaN     NaN      NaN   \n",
       "\n",
       "        yearsExperience  milesFromMetropolis         salary  \n",
       "count     999995.000000        999995.000000  999995.000000  \n",
       "unique              NaN                  NaN            NaN  \n",
       "top                 NaN                  NaN            NaN  \n",
       "freq                NaN                  NaN            NaN  \n",
       "mean          11.992407            49.529381     116.062398  \n",
       "std            7.212390            28.877721      38.717163  \n",
       "min            0.000000             0.000000      17.000000  \n",
       "25%            6.000000            25.000000      88.000000  \n",
       "50%           12.000000            50.000000     114.000000  \n",
       "75%           18.000000            75.000000     141.000000  \n",
       "max           24.000000            99.000000     301.000000  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_merge.describe(include = 'all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "#Numerical features correlation related to Salary\n",
    "train_merge_corr_indx = train_merge.corr().index\n",
    "plt.figure(figsize=(6,4))\n",
    "#plot heat map\n",
    "g=sns.heatmap(train_merge[train_merge_corr_indx].corr(),annot=True,cmap=\"Blues\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cat_boxplot(df, target, feature):\n",
    "    df_feature_mean = df.groupby([feature],as_index=False).mean()\n",
    "    df_feature_sorted = df_feature_mean.sort_values(by=target)\n",
    "    plt.figure()\n",
    "    sns.boxplot(x=feature, y=target, data=df, order=df_feature_sorted[feature].values)\n",
    "    plt.xticks(rotation=45)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "cat_features = ['companyId', 'jobType', 'degree', 'major', 'industry']\n",
    "target = 'salary'\n",
    "for feature in cat_features:\n",
    "    cat_boxplot(train_merge, target, feature)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "      <th>jobType_CEO</th>\n",
       "      <th>jobType_CFO</th>\n",
       "      <th>jobType_CTO</th>\n",
       "      <th>jobType_JANITOR</th>\n",
       "      <th>jobType_JUNIOR</th>\n",
       "      <th>jobType_MANAGER</th>\n",
       "      <th>jobType_SENIOR</th>\n",
       "      <th>jobType_VICE_PRESIDENT</th>\n",
       "      <th>...</th>\n",
       "      <th>major_MATH</th>\n",
       "      <th>major_NONE</th>\n",
       "      <th>major_PHYSICS</th>\n",
       "      <th>industry_AUTO</th>\n",
       "      <th>industry_EDUCATION</th>\n",
       "      <th>industry_FINANCE</th>\n",
       "      <th>industry_HEALTH</th>\n",
       "      <th>industry_OIL</th>\n",
       "      <th>industry_SERVICE</th>\n",
       "      <th>industry_WEB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>83</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>73</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10</td>\n",
       "      <td>38</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8</td>\n",
       "      <td>17</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>8</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   yearsExperience  milesFromMetropolis  jobType_CEO  jobType_CFO  \\\n",
       "0               10                   83            0            1   \n",
       "1                3                   73            1            0   \n",
       "2               10                   38            0            0   \n",
       "3                8                   17            0            0   \n",
       "4                8                   16            0            0   \n",
       "\n",
       "   jobType_CTO  jobType_JANITOR  jobType_JUNIOR  jobType_MANAGER  \\\n",
       "0            0                0               0                0   \n",
       "1            0                0               0                0   \n",
       "2            0                0               0                0   \n",
       "3            0                0               0                1   \n",
       "4            0                0               0                0   \n",
       "\n",
       "   jobType_SENIOR  jobType_VICE_PRESIDENT  ...  major_MATH  major_NONE  \\\n",
       "0               0                       0  ...           1           0   \n",
       "1               0                       0  ...           0           1   \n",
       "2               0                       1  ...           0           0   \n",
       "3               0                       0  ...           0           0   \n",
       "4               0                       1  ...           0           0   \n",
       "\n",
       "   major_PHYSICS  industry_AUTO  industry_EDUCATION  industry_FINANCE  \\\n",
       "0              0              0                   0                 0   \n",
       "1              0              0                   0                 0   \n",
       "2              1              0                   0                 0   \n",
       "3              0              1                   0                 0   \n",
       "4              1              0                   0                 1   \n",
       "\n",
       "   industry_HEALTH  industry_OIL  industry_SERVICE  industry_WEB  \n",
       "0                1             0                 0             0  \n",
       "1                0             0                 0             1  \n",
       "2                1             0                 0             0  \n",
       "3                0             0                 0             0  \n",
       "4                0             0                 0             0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cat_features.remove('companyId')\n",
    "num_features = ['yearsExperience','milesFromMetropolis']\n",
    "train_merge_Hot_Enc = train_merge[num_features].join(pd.get_dummies(train_merge[cat_features]))\n",
    "train_merge_Hot_Enc.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### <span style=\"color:black\">2.4.1 Observations</span>  \n",
    "> Years of Experience has a positve corellation, miles from metropolis has negative correlation, and cant be ignored.  \n",
    "> Company Id looks flat correlation, so can be ignored.  \n",
    "> All other features have some correlation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## <span style=\"color:black\">3. Develop</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">3.1. Fit Model</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#split the new df to training and test\n",
    "from sklearn.model_selection import train_test_split\n",
    "train_x, test_x, train_y, test_y = train_test_split(train_merge_Hot_Enc, train_merge['salary'], test_size=0.25, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "\n",
    "def cv(model, features, target,fold): \n",
    "    Rcross = cross_val_score(model,features,target, cv=fold, scoring='neg_mean_squared_error')\n",
    "    print(model)\n",
    "    print('Mean: '+str(- Rcross.mean())+', Std: '+str(Rcross.std()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">3.2. Linear Regression</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "lr = LinearRegression()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#fit the model\n",
    "#lr.fit(train_merge_Hot_Enc,train_merge['salary'])\n",
    "lr.fit(train_x,train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)\n",
      "Mean: 384.37444103478936, Std: 0.9150823851325356\n"
     ]
    }
   ],
   "source": [
    "#evaluate the model\n",
    "#cv(lr,train_merge_Hot_Enc,train_merge['salary'],2)\n",
    "cv(lr,train_x,train_y,2)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#calculate the prediction\n",
    "test_x_predict = lr.predict(test_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x6243181b48>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEICAYAAABWJCMKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3dd3hU1dbA4d8ihCpdRBQ0IFgQpHcJCAgICCqiYEVRUEQsVxS7n+VeVK4dQZqIIh00iEgRkYt0EZReBCWAVEF6Cfv7Y52ESUjCJEwyJet9njyZnDlzZk0mmXXOLmuLcw5jjDE5T65gB2CMMSY4LAEYY0wOZQnAGGNyKEsAxhiTQ1kCMMaYHMoSgDHG5FB+JQARaSUia0Vkg4j0SeX+vCIyxrt/oYjEeNtLiMgPInJQRD7y2b+AiEwRkTUislJE+gbqBRljjPFP7rPtICJRQH/geiAeWCwicc65VT67dQX+ds5VEJFOwJvA7cBR4EWgsvflq59z7gcRyQN8LyI3OOemphfL+eef72JiYvx8acYYYwB+/vnn3c65kim3nzUBAHWADc653wFEZDTQHvBNAO2BV7zb44GPREScc4eAuSJSwfeAzrnDwA/e7eMishQoc7ZAYmJiWLJkiR8hG2OMSSQif6S23Z8moIuBLT4/x3vbUt3HOXcS2A+U8DOwosCNwPf+7G+MMSYw/EkAksq2lPUj/NnnzAOL5AZGAR8kXmGksk83EVkiIkt27dp11mCNMcb4x58EEA+U9fm5DLAtrX28D/UiwF4/jj0IWO+cey+tHZxzg5xztZxztUqWPKMJyxhjTCb50wewGKgoIuWArUAn4I4U+8QB9wLzgVuBWe4sVeZE5HU0UTyQ0aCNMZlz4sQJ4uPjOXr0aLBDMVkgX758lClThujoaL/2P2sCcM6dFJGewDQgChjmnFspIq8CS5xzccBQ4HMR2YCe+XdKfLyIbAYKA3lE5CagBfAP8DywBlgqIgAfOeeG+P1KjTEZFh8fT6FChYiJicH7vzMRwjnHnj17iI+Pp1y5cn49xp8rAJxz3wLfptj2ks/to0DHNB4bk8Zh7a/PmGx29OhR+/CPUCJCiRIlyEhfqc0ENiaHsQ//yJXR99YSgAlbp07plzEmcywBmLCydy889xxcfjnkzw8FCkCVKvDUU7B+fbCjM/6IioqiWrVqVK5cmRtvvJF9+/Zl+lgxMTHs3r072bYuXbrwySefJNv21Vdf0bp16wwfK9JZAjBh49tvoXx56NsXKlaExx+Hnj2hTBl4/31NCk8/DSdPBjtSk578+fOzbNkyVqxYQfHixenfv39Aj9+5c2dGjx6dbNvo0aPp3LlzQJ8nElgCMGFh4kS46Sa47DJYvhymTIE334R+/WDqVPjzT+jWDd5+G5o3hwMHgh2x8Uf9+vXZunVr0s9vv/02tWvX5pprruHll19O2n7TTTdRs2ZNrr76agYNGpTuMZs3b86aNWvYvn07AIcPH2bmzJncdNNNfh1r8+bNVK58unRZv379eOWVVwDYuHEjrVq1ombNmjRq1Ig1a9YAMG7cOCpXrkzVqlWJjY3N3C8jCPwaBWRMMM2dC7fdBnXr6lVAkSJn7lO6NHzyCTRqBF26wF13waRJkMtOcdL2+OOwbFlgj1mtGryX5rzOZBISEvj+++/p2rUrANOnT2f9+vUsWrQI5xzt2rVjzpw5xMbGMmzYMIoXL86RI0eoXbs2HTp0oESJ1KvNREVFccsttzB27Fgee+wx4uLiuO666yhUqBBAho6VUrdu3Rg4cCAVK1Zk4cKF9OjRg1mzZvHqq68ybdo0Lr744nNq0spu9u9hQtquXdCpE5Qrl/aHv6+77tLPn7g4eOGF7InRZMyRI0eoVq0aJUqUYO/evVx//fWAJoDp06dTvXp1atSowZo1a1jvdex88MEHVK1alXr16rFly5ak7WnxbQZK2fyT0WMlOnjwIPPmzaNjx45Uq1aN7t27J11lNGzYkC5dujB48GASEhIy/DsJFrsCMCHLOT2b370b5s8/+4d/okcegV9/hf/8B9q2hQYNsjTM8OXnmXqgJfYB7N+/n7Zt29K/f3969eqFc45nn32W7t27J9t/9uzZzJw5k/nz51OgQAGaNGly1pnMDRs2ZPv27Sxfvpx58+YlJQN/jpU7d25O+QwvS7z/1KlTFC1alGWpXDUNHDiQhQsXMmXKFKpVq8ayZcv8vqoIJrsCMCFr7Fg963/zTahe3f/HicA778All2i/wPHjWRejybwiRYrwwQcf0K9fP06cOEHLli0ZNmwYBw8eBGDr1q3s3LmT/fv3U6xYMQoUKMCaNWtYsGDBWY8tItx2223ce++9tG7dmnz58gH4daxSpUqxc+dO9uzZw7Fjx/jmm28AKFy4MOXKlWPcuHGAzrxdvnw5oH0DdevW5dVXX+X8889ny5YtZxw3FFkCMCHpn3/giSegRg0d6ZNR550H/fvDypXw3/8GPj4TGNWrV6dq1aqMHj2aFi1acMcdd1C/fn2qVKnCrbfeyoEDB2jVqhUnT57kmmuu4cUXX6RevXp+Hbtz584sX76cTp2SKtP4dazo6Gheeukl6tatS9u2bbnyyiuT7hs5ciRDhw6latWqXH311Xz99dcA9O7dmypVqlC5cmViY2OpWrXqOf5msoecpWZbSKlVq5azBWFyhn/9C959FxYsgDp1Mn+cW26B6dNh40YoVSpw8YWr1atXc9VVVwU7DJOFUnuPReRn51ytlPvaFYAJOfHx8NFH2v5/Lh/+oM1Hx46BN4rPGOPDEoAJOW+8oR3APsPAM61iRXjoIRg8GLwh28YYjyUAE1I2bYIhQ+DBB+HSSwNzzBdf1LIRgUgoxkQSSwAmpPTtC1FRWu8nUC64AB59FMaNg9WrA3dcY8KdJQATMrZvh+HD4b774OKLA3vsJ5/UwnFvvBHY4xoTziwBmJDx3ntayO2ppwJ/7PPPhx49YNQoWLcu8Mc3JhxZAjAhYd8+GDBAa/5cdlnWPMe//gW5c2vlUBM8vuWgO3bsyOHDhzN9rNmzZ9O2bVsA4uLi6Nu3b5r77tu3j48//jjDz/HKK6/Qr1+/M563fv36ybadPHmSUqVKJZWH8PdYwWQJwISEAQO0guczz2Tdc5QqBXfeqc1Mf/+ddc9j0udbDjpPnjwMHDgw2f3OuWSlGPzVrl07+vTpk+b9mU0AqYmNjSU+Pp7NmzcnbZs5cyaVK1emdOnSAXmO7GAJwATd0aN6Vt6ypRaTzEqPPQaHD+tIIxN8jRo1YsOGDWzevJmrrrqKHj16UKNGDbZs2cL06dOpX78+NWrUoGPHjkklIr777juuvPJKrr32WiZOnJh0rOHDh9PTmza+Y8cObr75ZqpWrUrVqlWZN28effr0YePGjVSrVo3evXsDaZeffuONN7jiiito3rw5a9euPSPuXLly0bFjR8aMGZO0zbfo3ODBg6lduzZVq1alQ4cOqV7lNGnShMSJrbt37yYmJgbQKqm9e/dOiitxcZvt27cTGxubdPX0v//9L9O/90RWDM4E3WefwY4dWXv2n6hqVbjuOvjwQy01kTsH/wcEuRo0J0+eZOrUqbRq1QqAtWvX8umnn/Lxxx+ze/duXn/9dWbOnEnBggV58803eeedd3j66ad58MEHmTVrFhUqVOD2229P9di9evWicePGTJo0iYSEBA4ePEjfvn1ZsWJFUjG3tMpPFyxYkNGjR/PLL79w8uRJatSoQc2aNc94js6dO9OtWzeeeeYZjh07xrfffsu7774LwC233MKDDz4IwAsvvMDQoUN59NFH/fq9DB06lCJFirB48WKOHTtGw4YNadGiBRMnTqRly5Y8//zzJCQknFPTWaIc/OdvQkFCgi7qUrs2NGmSPc/5+OPQvr2uF9CxY/Y8pzktsRw06BVA165d2bZtG5deemlSbZ4FCxawatUqGjZsCMDx48epX78+a9asoVy5clSsWBGAu+66K9VFXWbNmsWIESMA7XMoUqQIf6do9/MtPw1a7nn9+vUcOHCAm2++mQIFCgDatJSa2rVrc/DgQdauXcvq1aupV68exYoVA2DFihW88MIL7Nu3j4MHD9KyZUu/fz/Tp0/n119/Zfz48YAWsFu/fj21a9fm/vvv58SJE9x0001Jv8NzYQnABNXEibBhA4wfr1U8s0ObNtrR/N57OTsBBKkadFIfQEoFCxZMuu2c4/rrr2fUqFHJ9lm2bBkSoD+UtMpPv/fee34/R6dOnRg9ejSrV69OtuZAly5d+Oqrr6hatSrDhw9n9uzZZzzWt+y0b0lq5xwffvhhqkljzpw5TJkyhbvvvpvevXtzzz33+BVnWqwPwASNc1qrp2JFXe4xu0RFQa9eMG8eLFqUfc9r/FevXj1++uknNmzYAOiyjuvWrePKK69k06ZNbNy4EeCMBJGoWbNmDBgwANA29X/++YdChQpxwGet0LTKT8fGxjJp0iSOHDnCgQMHmDx5cppxdu7cmS+++IJZs2Ylu1I4cOAApUuX5sSJE4wcOTLVx8bExPDzzz8DJJ3tJ8Y1YMAATpw4AcC6des4dOgQf/zxBxdccAEPPvggXbt2ZenSpen/Ev1gCcAEzaxZ8PPP0LvJYqJGjoAvv4QlSzQzZLH77oPChYN3FmzSV7JkSYYPH07nzp255pprqFevHmvWrCFfvnwMGjSINm3acO2113JpGvVC3n//fX744QeqVKlCzZo1WblyJSVKlKBhw4ZUrlyZ3r17p1l+ukaNGtx+++1Uq1aNDh060KhRozTjrFSpEgUKFKBp06bJrmBee+016taty/XXX5+snLSvp556igEDBtCgQQN2796dtP2BBx6gUqVK1KhRg8qVK9O9e3dOnjzJ7NmzqVatGtWrV2fChAk89thjmfztnmbloE1wLF9Oi+sdv+0qxSbKkY9jp++7+GK49VadEVamTJaF8OST2hn8xx9w0UVZ9jQhxcpBRz4rB21C25QpLK3Xgxm7qvH4tT+Tb8GPWrB/9WoYMUJrQH/8MVSooEN1smjQfs+e2gnttRQYk+P4lQBEpJWIrBWRDSJyxkwLEckrImO8+xeKSIy3vYSI/CAiB0XkoxSPqSkiv3mP+UAC1bNjQtuQIdCuHW/lf4nChU7x0DdtoW5dKF8errwS7r5be4bXrdNZWx9+CFWqwIwZAQ+lfHm48Ub45BOdi2BMTnPWBCAiUUB/4AagEtBZRCql2K0r8LdzrgLwLvCmt/0o8CKQWnWXAUA3oKL31SozL8CEkfnz4aGH2NjwHsbtb8FDD+dKe6H3mBgYOlSXBCtUCFq00CahhISAhtSrF+zaBT7zeSJeODX7mozJ6HvrzxVAHWCDc+5359xxYDTQPsU+7YHPvNvjgWYiIs65Q865uWgiSCIipYHCzrn5TiMeAWTjOBCT7fbvhzvugLJleffKgeTOLTz+uB+Pq1ULli7VSm7//a8OF/JGbQRC06Zw9dW6AllOkC9fPvbs2WNJIAI559izZw/58uXz+zH+zAO4GPBd4j4eqJvWPs65kyKyHygB7CZ1F3vH8T1mgAsAm5DhHDz8MGzZwv6p8xh+c17uuAP8LpmSP7+u8F6pkp6yN2oEkycHpINYBLp318P+8gt4c4IiVpkyZYiPj2fXrl3BDsVkgXz58lEmA/8X/iSA1NrmU54++LNPpvYXkW5oUxGXXHJJOoc0Ievbb7UO86uvMnxVHQ4d0gVaMuyRR7Th/rbbtN9g8mSoUeOcw7v7bi1D8cknkKIuWcSJjo6mXLlywQ7DhAh/moDigbI+P5cBtqW1j4jkBooAe89yTN80ldoxAXDODXLO1XLO1SpZsqQf4ZqQcvIkPP00VKjAqd7P8NFHUL/+OXxu33AD/PSTzuZq1AhmzjznEIsWhdtvh5EjtSKpMTmFPwlgMVBRRMqJSB6gExCXYp844F7v9q3ALJdOI6NzbjtwQETqeaN/7gG+znD0JvQNHw6rVkHfvkyfnYcNGzJ59u/rmmt0Cu9ll0G7dpDKNPuM6tZNuxZyUmewMX5NBBOR1sB7QBQwzDn3hoi8CixxzsWJSD7gc6A6eubfyTn3u/fYzUBhIA+wD2jhnFslIrWA4UB+YCrwaHpJA2wiWNg5dEjrPMTEwE8/cUsHYe5ciI+HPHkCcPydO7W05x9/wLRp4BUOywzndBRq2bIBuagwJqSkNRHMr2JwzrlvgW9TbHvJ5/ZRINWyWs65mDS2LwEq+/P8Jkx98IEu9Dt+PDt2CpMnayXOgHz4g672/v332hTUqZNeaRQqlKlDiWjXwr//rXnlggsCFKMxIcxmApuscfgwvPOOttk3aMCIEdod0LVrgJ/nwgt19nB8PPgs6JEZt90Gp07pPDRjcgJLACZrDBkCu3fD88/jnP547bXazBJw9evrWM7339exnJlUubLGN3ZsAGMzJoRZAjCBd/w4vP02xMZCw4bMn6+VHQJ+9u/rP/+BkiW1NzcT68nC6WagH3+Ev/4KcHzGhCBLACbwvvhCm2Seew7QM+q8eeGWW7LwOYsVg759tZz0d99l+jAdOmj+mDo1gLEZE6IsAZjASkjQD+IaNaBFC06dgnHjtCugcOEsfu4779RS0v/9b6YPUaWKditMnx7AuIwJUZYATGBNmADr1+vZvwjz58O2bdm09GJ0NDz2mK40k8m+ABGtOzdjRqZbkowJG5YATOA4p+Mor7wSbr4ZON38c+ON2RTDgw/Ceeed01VAixawZ4/WoDMmklkCMIEzdSosXw59+kCuXJw6pYu9t26d6eH5GVe0qCaB0aNhy5az75+K66/X79YMZCKdJQATGM7BG2/AJZdo2Wd00fVsa/7x9dhj2n4zaFCmHn7BBVoV1BKAiXSWAExgzJ+vn/i9e2tbPNr8ky8ftG2bzbFceim0agXDhunss0xo0UJrzgVw6QFjQo4lABMYH30ERYpAly6ADgbK9uYfX9266eVHJsdzNmmiuWPBgsCGZUwosQRgzt327TrW8777tAMWPXvevj0IzT+J2rTR8ZyZbAZq0ABy5YL//S/AcRkTQiwBmHM3aJCeLj/ySNKmceOC1PyTKDpaE9K33+qktAwqXBiqVrUEYCKbJQBzbo4f12W0brgBKlQATjf/tGmTdEEQHF27amfwp59m6uHXXqtNQMePBzguY0KEJQBzbiZO1MI5Pqu8zJ2rm4LW/JPosst0TGfiFUoGNWoER46cU305Y0KaJQBzbgYPhnLloGXLpE1jx+o67m3aBDGuRD17ahPQV19l+KGNGul3awYykcoSgMm8jRu17ML992uPKdr8M2FCCDT/JGrTRlck+/DDDD/0wgu1VcsSgIlUlgBM5n36qX7we0M/QT8sd+zQssohISpKrwLmzNFZyhl07bU6osmPlVONCTuWAEzmJCTogu+tWkGZMkmbx4zR5p/WrYMX2hnuvx8KFMjUVUCDBloXaP36LIjLmCCzBGAyZ9o02Lo12Sovx49r+/9NN0HBgkGMLaVixeCuu2DkSF3wNwMaNNDv8+ZlQVzGBJklAJM5Q4boClw+A/2/+w727tXP2pDz5JNw7JguVJ8BV12lE5wtAZhIZAnAZNyOHTB5MtxzD+TJk7T5iy/g/PNPV9MMKVdcoct9ffQR/POP3w/LlUuXHJ4/PwtjMyZILAGYjPv8cx1X79P8s38/xMVBp05JteBCz7PPaqADBmToYfXrw8qVsG9fFsVlTJBYAjAZ45w2/zRooO0jnokTtYUlJJt/EtWoofMV3nlHZ3j5qUEDfdkLF2ZhbMYEgSUAkzHz5sHatcnO/kGbfypUgDp1ghSXv559VjuCP/vM74fUqaNNQdYMZCKNJQCTMUOH6gwvn4H+8fHwww+6JrtIEGPzR2ws1KoF77/v96K/hQtD5crWEWwijyUA478DB3ScZ6dOyab5jhqlTSR33hnE2PwlAo8/DmvW6MrvfmrQQAvDJSRkYWzGZDO/EoCItBKRtSKyQUT6pHJ/XhEZ492/UERifO571tu+VkRa+mx/QkRWisgKERklIvkC8YJMFpo4EQ4d0jLLPkaOhLp1oWLFIMWVUR07QunS8N57fj+kQQPNfytXZmFcxmSzsyYAEYkC+gM3AJWAziJSKcVuXYG/nXMVgHeBN73HVgI6AVcDrYCPRSRKRC4GegG1nHOVgShvPxPKRo6E8uV1WIxn5UqtsBAWZ/+J8uSBHj104sLq1X49JHFCmPUDmEjizxVAHWCDc+5359xxYDTQPsU+7YHEXrXxQDMREW/7aOfcMefcJmCDdzyA3EB+EckNFAC2ndtLMVlq2zb4/vszGvrHjNEO0qCXfs6o7t0hb16/J4aVL6/z3qwfwEQSfxLAxcAWn5/jvW2p7uOcOwnsB0qk9Vjn3FagH/AnsB3Y75ybntqTi0g3EVkiIkt27drlR7gmS4werZ2mPqf6zmmXQOPGWjkzrJQsqX0ZX3zh18rvInoVYAnARBJ/EkBq4zpS1kZMa59Ut4tIMfTqoBxwEVBQRFIdQe6cG+Scq+Wcq1WyZEk/wjVZYuRIHT1zxRVJm379VUeEhkzlz4zq3l0//EeP9mv3Bg1gwwaw8xATKfxJAPFAWZ+fy3Bmc03SPl6TThFgbzqPbQ5scs7tcs6dACYCDTLzAkw2WLUKli49o6F/7FitttyhQ5DiOlf16un4zk8+8Wt36wcwkcafBLAYqCgi5UQkD9pZG5dinzjgXu/2rcAs55zztnfyRgmVAyoCi9Cmn3oiUsDrK2gG+NcbZ7Lfl19qQ3+n0/30zmn7f9Om2poSlkT0KmDJEk1wZ1GzJuTObc1AJnKcNQF4bfo9gWnoh/RY59xKEXlVRNp5uw0FSojIBuBJoI/32JXAWGAV8B3wiHMuwTm3EO0sXgr85sUxKKCvzARGYkP/ddcla+j/5RddECxsm38S3XUX5Mun6wafRf78Wk3CEoCJFOLCaKmjWrVquSVLlgQ7jJxl2TKoXl2bSbp1S9r8zDNaUuevv6BEiSDGFwhduug6ltu3n3UdyyeegIEDtaBoyBa9MyYFEfnZOVcr5XabCWzSN2aMNvTfckvSpsSLgubNI+DDHzLUGdygARw9qnnRmHBnCcCkLfGTvlkzLfTvWbwYNm+G228PXmgBlYHO4MQ5cNYMZCKBJQCTtp9/ht9/P+OTfuxYbf5on3I6YLjKQGdwmTJQtqyNBDKRwRKASdvYsTrs5aabkjadOqWbW7bUpXYjRmJn8ODBZ93VJoSZSGEJwKTOORg3Ttd3LF48afPChbBlSwSM/kmpaFG90hk58qwzgxs00N/Bli3p7mZMyLMEYFK3bJk29KeY5TVmjJbQiZjmH1/du2vJz7N0Bif2A1gzkAl3lgBM6iZN0slf7dolbTp1Si8KWrXSRVIiTr16uszlsGHp7latms4JsARgwp0lAJO6iRN19Syfab4//aRFQSNm9E9KIrrU5fz56ZaJjo6G2rWtH8CEP0sA5kxr12qh/5tvTrZ5zBjtJ23bNkhxZYe779aO77NcBTRooAOGMrC2vDEhxxKAOdOkSfrdJwEkJMD48dCmDRQqFKS4ssMFF8CNN8KIEXDiRJq71a8PJ0/qSFljwpUlAHOmSZO0jaPs6UKuc+bAjh0ROPonNfffDzt3wpQpae5iE8JMJLAEYJKLj4dFi5KVfgAd+1+ggF4BRLxWrXTN4HSagUqW1DWQLQGYcGYJwCSXSvPPyZNaK61tWyhYMEhxZafcueHee+Hbb7VAXBoSJ4SFUT1FY5KxBGCSmzQJKlVKtvLXjz/qKlgRO/onNfffrx0fI0akuUvDhvp7Wb8+G+MyJoAsAZjTdu/WT/sUo3++/lpH/7RqFaS4gqFiRWjUSJuB0jjFj43V73PmZGNcxgSQJQBz2uTJOtsrRennyZO19HOBAkGMLRi6doV163QCRCouv1wHDVkCMOHKEoA5beJEuPRSXQDGs2KFVoTwmRCcc9x6q455HTo01btF9CrAEoAJV5YAjDpwAGbM0OYfkaTNkyfr94ie/JWWggV1HeSxY/X3k4rYWPjjD/0yJtxYAjBq6lQ4duyM9v+4OJ0SULp0kOIKtvvvh8OHdRp0KhL7Af73v2yMyZgAsQRg1KRJOri9YcOkTX/9pVMCcmTzT6K6dXVUVBpzAipX1krS1gxkwpElAKNn/lOmaI3nqKikzVOmaCfwjTcGMbZgE9GrgDQKxEVFwbXXWgIw4ckSgIHvv9c27hSzfydPhksugWuuCVJcoeIsBeJiY7V+3o4d2RyXMefIEoDR0T+FCkHTpkmbjhyB6dP17N+nTzhnOkuBOOsHMOHKEkBOl5CgM73attWlvjyzZmkSyNHt/766dtUCcXFxZ9xVo4bOkbBmIBNuLAHkdHPn6gzgVEb/nHceNG4cpLhCTatWOkfiww/PuCs6WusCWQIw4cYSQE43aZKe+d9wQ9KmU6fgm2/0M8/noiBni4qCnj21VMayZWfcHRsLv/4K+/YFITZjMskSQE526pSW+WzRQk/3PUuX6tKPOXr0T2q6dtW2ng8+OOOu2FgdMZVG1QhjQpJfCUBEWonIWhHZICJ9Urk/r4iM8e5fKCIxPvc9621fKyItfbYXFZHxIrJGRFaLSP1AvCCTAXPmaP3/zp2TbZ48WdeDb906SHGFqmLFtEz0l19qGVAfdepAnjx6gWBMuDhrAhCRKKA/cANQCegsIpVS7NYV+Ns5VwF4F3jTe2wloBNwNdAK+Ng7HsD7wHfOuSuBqkDaq3CbrDFypJY7SNHTGxenbdrnnx+kuEJZr146b2LgwGSb8+eHevXghx+CFJcxmeDPFUAdYINz7nfn3HFgNNA+xT7tgc+82+OBZiIi3vbRzrljzrlNwAagjogUBmKBoQDOuePOOWs9zU7HjukivzffnGyVly1btInbRv+k4cortXPko4/OWBG+aVNtPvv77yDFZkwG+ZMALga2+Pwc721LdR/n3ElgP1AinceWB3YBn4rILyIyRERSXWtKRLqJyBIRWbIrxWW3OQdTp2qP5Z13JtucWPzN2v/T8fTTOiT0s8+SbW7aVLtVbDSQCRf+JIDUpgGlXCEjrX3S2p4bqAEMcJEjY4IAACAASURBVM5VBw4BZ/QtADjnBjnnajnnapUsWdKPcI1fRo7UCU7NmyfbPHmyroXisyCYSalJE23079dP51F46tTRpiBrBjLhwp8EEA+U9fm5DLAtrX1EJDdQBNibzmPjgXjn3EJv+3g0IZjssH+/ftLffruWOPAcOKATwNq1s9m/6RKBZ56BjRt1FJUnb16tCzRrVhBjMyYD/EkAi4GKIlJORPKgnbopp0PGAfd6t28FZjnnnLe9kzdKqBxQEVjknPsL2CIiieeZzYBV5/hajL8mTtQ+gBTNPzNmwPHj1vzjl/btdUmwN99MtmRk06bw22/aQmRMqDtrAvDa9HsC09CROmOdcytF5FURSewqHAqUEJENwJN4zTnOuZXAWPTD/TvgEedc4jXzo8BIEfkVqAb8O3Avy6Rr5Ei47DJts/ARF6cjHX0qQpu0REXBU09pr6/P4P/rrtPvs2cHJyxjMkJcGgteh6JatWq5JUuWBDuM8LZtG5QpAy+8AK++mrQ5IQEuvBBatoQvvghifOHk8GG4+GL9pY0eDcDJk1C8uF5cDRgQ5PiM8YjIz865Wim320zgnGb0aG2ySNH8s2CBlgSy4Z8ZUKCArhUwYYImVrRLpXFj6wcw4cESQE4zciTUrHnGMJ9Jk7SoWcuWaTzOpO7hh/XyadCgpE3XXQfr1ukka2NCmSWAnGTNGm2zTnH275yue96yJRQpEqTYwlWFCjox7JNPtAed08sq2HBQE+osAeQko0bpEMZOnZJtXrBAZwDfdluQ4gp3PXvqAspffQXoCmrFi1sCMKHPEkBOMmECNGoEpUsn2zxmjI5hb5+ywIfxT+JaAUOGAFpIr0kTXWkzjMZYmBzIEkBOsXYtrFwJHTok23zqFIwbp8sBFC4cpNjCXa5c2hk8YwZs2gRoM9Cffyb9aExIsgSQU0ycqN9TrPw1Z44OYLn99iDEFEnuu0+b17yF4xP7AWw0kAlllgByiokTdeJX2bLJNn/6qZ752/DPc1S2rDYFffopnDzJlVfqvArrBzChzBJATvDHH7BkyRnNP/v3a/NP5846pN2cowcegK1bYdo0RHQ46KxZ1g9gQpclgJxg0iT9fsstyTaPGaMl7bt2DUJMkejGG7XCqtcZ3LSpDg5asybIcRmTBksAOcGECTo2sUKFZJuHDoXKlaHWGRPETaZER0OXLlpp9a+/kuoCWT+ACVWWACLdX39psbIUZ/8rVsCiRTp4xUo/B1DXrjoz+LPPKF8eLrnE+gFM6LIEEOm++koboVO0/w8bpiesd90VpLgi1eWXQ2wsDBmC4GjaVBPAqVPBDsyYM1kCiHQTJ+oSX1dfnbTp+HH4/HMd+WOLrGWBBx6ADRvgxx9p3Bj27tUpGMaEGksAkWzvXj397NAhWTvP5Mla+dM6f7NIhw5aVGnIEBo31k22TrAJRZYAItnkyVqgPkX7/7BhWsa+RYsgxRXpChTQtrXx44kp8jdly8KPPwY7KGPOZAkgkk2YoL2QPsN8du2CadPg7rt1USuTRR54AI4dQ74cSWysXgHYfAATaiwBRKoDB2D6dD3792n+GTdOB6nccUcQY8sJqlXTdRcGD6ZxrGPHDl0jwJhQYgkgUn37rS78nqL5Z9Qo7Q+uUiVIceUkDzwAv/5KbPEVgDUDmdBjCSBSTZgApUpBgwZJm/74A+bOtbP/bNO5M+TPz+Uz+lOqlHUEm9BjCSASHTmiVwA33ZSsod9btzzlejAmqxQpArfdhoz6ksYNT/Djj9YPYEKLJYBINH06HDp0xuSvUaOgXj0oXz5IceVEDz4IBw4Qm28x8fG2PoAJLZYAItHEiVCsmC5L5Vm1CpYvt+afbNegAVx9NY2XvQ9YM5AJLZYAIs3x4xAXp9N8o6OTNo8apQtX2bq/2UwEevSg0qpxlChywjqCTUixBBBpZs+GffuSjf5xDr78Epo1035hk83uuotc5xWkUZFfLQGYkGIJINKMHQvnnZdsmu/ixfD77zooxQRB4cJw11003jaaTZtgy5ZgB2SMsgQQSY4d0+GfN98M+fIlbf7yS8ib94wpASY7PfwwjU/OBKwfwIQOvxKAiLQSkbUiskFE+qRyf14RGePdv1BEYnzue9bbvlZEWqZ4XJSI/CIi35zrCzFojYd9+5Kd6ick6MpfrVvrqEQTJNdcwzXXFqForv3M+t5qQ5vQcNYEICJRQH/gBqAS0FlEKqXYrSvwt3OuAvAu8Kb32EpAJ+BqoBXwsXe8RI8Bq8/1RRjPl19CiRLQvHnSptmzdU0YG/0TfFFP9KLZqRlMjztq8wFMSPDnCqAOsME597tz7jgwGmifYp/2wGfe7fFAMxERb/to59wx59wmYIN3PESkDNAGGHLuL8Nw8KCO/unY8YzRP4UKQZs2QYzNqPbtaVFiKfF7Ctg6wSYk+JMALgZ8u63ivW2p7uOcOwnsB0qc5bHvAU8D6V4Pi0g3EVkiIkt27drlR7g5VFyczgD2OdU/dgzGj9cugfz5gxibUVFRtOih6zJPH7w5uLEYg38JILUVY1NewKa1T6rbRaQtsNM59/PZntw5N8g5V8s5V6ukLV+VtlGjoEwZaNgwadM338D+/Tb6J5TE9O7I5bnWM3303mCHYoxfCSAeKOvzcxlgW1r7iEhuoAiwN53HNgTaichmtEmpqYh8kYn4DcCOHTB1qp795zr9ln76KVx0EVx/fRBjM8kVKkSLqjuYvf0Kjq3aGOxoTA7nTwJYDFQUkXIikgft1I1LsU8ccK93+1ZglnPOeds7eaOEygEVgUXOuWedc2WcczHe8WY552x58swaOVKH+3TpkrRp+3bNCffcYwu/hJoWva7iMAWZ+9SkYIdicrizJgCvTb8nMA0dsTPWObdSRF4VkXbebkOBEiKyAXgS6OM9diUwFlgFfAc84pxLCPzLyMGc01P9unXhqquSNn/+OZw6BffdF8TYTKquu7UEeaNO8M130bB5c7DDMTmYuDAaj1arVi23ZMmSYIcRWpYu1ZWnBgyAhx4CNCdUqgTFi8NPPwU5PpOqts2PsOr7v9j4YF9k0CfBDsdEOBH52TlXK+V2mwkc7oYP12m+t9+etGnOHFizBrp2DV5YJn3tbsvPJsqx8tNF8OefwQ7H5FCWAMLZsWPa/n/zzVr+2dO/v/5oC7+ErrZt9XucawuvvRbcYEyOZQkgnH31Fezdm6zzd+tWXQ6ga1coUCB4oZn0XXQR1K4Nced31T4cWzHeBIElgHA2YIAu7+UzznPQIO38ffjhIMZl/NK+PSzcEcP2PJfCyy8HOxyTA1kCCFerV8OPP0L37klj/48ehYEDtfCbLfsY+tp5Y+i+ue6/umDz8uXBDcjkOJYAwtUnn2jNH59xniNGwM6d8NRTQYzL+K1yZYiJgbiENlC0KDz/fLBDMjmMJYBwdPgwfPYZ3HoreOUxEhKgXz+oVQsaNw5yfMYvItoMNPPHaA498QJMmWLjdk22sgQQjsaO1br/3rh/0Fpw69fD00/rB4sJD+3aadPdjMsfgQsvhOeew2pFm+xiCSAcDRyos34bNQL08+LNN7Xd31b9Ci+NGmnrT9z0fPDCCzqJY/r0YIdlcghLAOHml19g4UI9+/dO9efO1U3/+pfV/Qk30dHaaf/NN5Bw/4PaKWBXASabWAIINwMHanH/u+9O2vT223D++cmmA5gw0q4d7NoFC3/JA6+8ouU9Jk4MdlgmB7AEEE7++Udn/nbqlDTzd9UqmDwZeva0iV/hqlUryJ0bvv4auOsubd574QXt2TcmC1kCCCcjR8KhQ8k6f995Ry8IHnkkiHGZc1KkCDRpoh35REXB669rMafPPw92aCbCWQIIF87pzN/q1bWGALrY++efa9PP+ecHNzxzbtq318/8devQ2k41a2pz0LFjwQ7NRDBLAOFi7lz47Tfo0SOp87d/fzhxAp54IsixmXN24436ffJk9P3997/hjz9g8OCgxmUim60HEC5uv12HB27dCgUKcPgwXHIJXHut1oQz4a9aNTjvPM31OAdNm2rJj40boWDBYIdnwpitBxDOtm3TUSH335/U0/vZZ7Bnjw79NJHhlltg3jxt2kME3nhD13v+4INgh2YilCWAcDBokI4I8Up8JiTAu+9CnTp6BWAiQ4cOeuKfdEXXoIEuHPDWW/D330GNzUQmSwCh7vhxLfx2ww1QoQKg7cTr1+vZv5V9iByVKsHll6eYAvD661r24623ghaXiVyWAELdpEnaJuAzzvO//4VLL7WyD5FGRN/TH37QdX4AqFpVJ/29845mfWMCyBJAqPvoIy3y06oVAAsWaCfh44/r5CETWW65BU6e9OYEJHrrLciXT2f7hdGgDRP6LAGEsuXL9dO+R4+kRV/69dPiYQ88EOTYTJaoVQvKlYNRo3w2Xnihrhs8fbqViDABZQkglPXvr9N8vUVfNmzQ//+HH9bhgibyiMAdd8DMmd5ooEQ9emhz0OOPw/79QYvPRBZLAKHq77/hiy/gzjuheHFAm4Gjo+HRR4Mcm8lSd96p6zqPHu2zMXduHQywbZsmAWMCwBJAqBo2DI4cSer83bULPv1U+wNLlw5ybCZLXXUV1KihpZ+SqVsXnn0Whg/3KscZc24sAYSiEyfg/fchNlanhwIff6wrR9nEr5zhrrtgyRKtD5TMSy/p38SDD+oC0MacA78SgIi0EpG1IrJBRPqkcn9eERnj3b9QRGJ87nvW275WRFp628qKyA8islpEVorIY4F6QRFhzBjYsgV69wZ0CeCPPtI5QVddFeTYTLa44w5t9TmjFFCePFoBcP9+rQpro4LMOThrAhCRKKA/cANQCegsIpVS7NYV+Ns5VwF4F3jTe2wloBNwNdAK+Ng73kngX865q4B6wCOpHDNnck5XeLnqKl0qCi37sHt3Uj4wOUCpUnDTTdrac/RoijsrV9YyEZMmWcloc078uQKoA2xwzv3unDsOjAbap9inPfCZd3s80ExExNs+2jl3zDm3CdgA1HHObXfOLQVwzh0AVgMXn/vLiQAzZsCvv+qnfa5cHD8OfftCvXpJSwCbHOKhh3RC2Pjxqdz5xBP6B/Hoo/Dnn9kem4kM/iSAi4EtPj/Hc+aHddI+zrmTwH6ghD+P9ZqLqgML/Q87gr39tvby3nEHoGf/f/4JL79sZR9ymuuu0+ofn3ySyp1RUXp5kJCgC0LY6mEmE/xJAKl97KRseExrn3QfKyLnAROAx51z/6T65CLdRGSJiCzZtWuXH+GGsV9+0QHgjz0GefNy4oSWha9TB1q2DHZwJrvlyqVXAXPn6jLBZyhfXiuF/vCDzhA0JoP8SQDxQFmfn8sA29LaR0RyA0WAvek9VkSi0Q//kc65NKc3OucGOedqOedqlSxZ0o9ww1i/fjrDq3t3QJt3N2/WgR929p8zPfAAFCqkF4apuu8+6NhR1xBevDhbYzPhz58EsBioKCLlRCQP2qkbl2KfOOBe7/atwCynK83EAZ28UULlgIrAIq9/YCiw2jn3TiBeSNj74w8d/dOtGxQtyokT2s9Xs2ZSX7DJgYoU0fOBceP0ZOAMItpGlNhseOBAdodowthZE4DXpt8TmIZ21o51zq0UkVdFpJ2321CghIhsAJ4E+niPXQmMBVYB3wGPOOcSgIbA3UBTEVnmfeXsj7l339V/Zm+W58iR8PvvdvZvtEVQRP9EUlWs2Ok/mF69sjU2E95sSchQsHevru94883w+eecPKmjQAsVgp9/tgRg4N57dTTQn39CiRJp7PTii7p+wOjRuoSoMR5bEjKUffABHDqUNND/iy+08Jud/ZtETz2lEwIHDEhnp5de0vHC3bun0V5kTHJ2BRBs//yjq7s0bgxffcWJE3DFFVr/bfFiSwDmtNat9Yrwjz90eYBU/f67FhIqXx5++kmryZocz64AQtVHH+mSfy++CGjBt02b4NVX7cPfJNe7t5b/+eyzdHYqX177A5Yt0wEFYXSCZ7KfXQEE08GDEBOjA/2//ZajR6FiRShTBubNswRgknNOC4Lu2QNr155lRbjXXtMmoXfe0VnDJkezK4BQNHCg/jd7Z/+DB0N8vP7v2oe/SUlEh/v//nuKtQJS8/zzur7kv/6lw4uNSYVdAQTLkSO69l/lyjBzJocPw2WXweWXw+zZlgBM6k6dgurV4fhxWLFCK0Kk6cgRnUK+YAFMmQLXX59tcZrQYlcAoWbIENixI+nsf8AAXQLQzv5NenLl0pP7NWv8WB44f35dXf6qq3SI8f/+ly0xmvBhVwDBcOyYnu6XLw9z5rBzp575160L06YFOzgT6hIS9MIxTx7t6z3rCcNff2lluS1b4NtvdaEhk6PYFUAoGT4ctm5NOvt/7jmdBvDee8ENy4SHqCj9m/n1V5g82Y8HXHihFowrW1bHks6aleUxmvBgCSC7HTumJT7r1oXmzVm0SJf/7dXLVvsy/uvcWS8gX3/dz5GeiUkgJgZuuEEXkzE5niWA7DZwoM7nf/11Dh8R7r1X63i9/HKwAzPhJHduXR9+8WJt1fHLhRfCnDk6UezWW2HQoCyN0YQ+SwDZ6cABPWVr1gyaN+fpp7Uzb/hwKFw42MGZcHPvvdqV9PzzOjrIL8WL65oTLVpoyYjHH4eTJ7M0ThO6LAFkp3fe0cV9//Mfxo+H/v11jo6NzjOZER2tM8aXL9dy0X4rWFA7Dx57DN5/X/sFIn2xJZMqGwWUXXbu1NO1li1Z/Mx4YmP1SnzWLMibN9jBmXB16hRUq6ZD/lesyMTf0pAh8MgjcP75OrvMFp6OSDYKKNieew6OHuX3h96iXTttjp00yT78zbnJlQveekurx76TmaWVHnhAJ4oVKABNmuhU42PHAh2mCVGWALLDkiUwbBhb73+R5t3Kc/y4Tsy84IJgB2YiQatWWvXhtde0UmiGVa+uZUbvvvv0MnTheqVtMsQSQFY7dQp69eKvEldz/Y/Ps2sXfPcdVKoU7MBMJHnvPZ0Q1qNHJguAFi6soxGmTNHqtPXq6TCjo0cDHaoJIZYAstrnn7N9/iaa5J7Ln/FRTJkCtWsHOygTacqWhb59dUhoppqCErVurZ0J996rB6xe3SaORTBLAFlp61a29nqTJvkWsvVgYaZOtVn4Juv07Kklf/r00XLimVa0KAwdClOnan9As2Zw221ahtREFEsAWcU5ttz9HNcd+JptUWX47juxARYmS4norPJLL4V27WD16nM8YKtWsHIl/N//6bDRK67QRWZsucmIYQkgiyx++Rvq/NCXHXkvYdr0XDRsGOyITE5QtKj2MeXOrfNLNm06xwPmz68Ly2zcCA89pMuRXXaZ9jr/+KOtOBbmLAEEmHMw7NV4Gr/WjHz5YN6iaBo0CHZUJiepUAGmT9dF5K+9FlatCsBBL7oIPvxQE8Ezz2hJiSZNdBLC0KE6EcGEHUsAAfTPP3Dnrcfo+nIZ6udZysJFubi6iv2KTfa75ho9QT91Sud2BWwpgDJltJjhli36wQ86l6BMGR019OefAXoikx3s0ylAFi+GGjUcYyfl5vWol5j+QzQXVCkV7LBMDlalCsydq5N8mzWDTz8N4MHz54f779cFCWbP1quBt97SVe7atdPhSAkJAXxCkxUsAZyjAwe0nk+9eo7j23bzo4vl+cHliGpQN9ihGcNll+lE39hY/by+5x79mw0YEWjcGCZM0FFCzz4LixZBmzb65P/+ty5IY0KSJYBz8NVXOqHr/fcd3WOm8+uRijR8tyPcd1+wQzMmSbFi2jH8yiswcqT+zY4ZkwX9t5deqtVut2zR6nQVKmip0rJldRjplCk2sSzEWALIIOd0XY0mTXTMdbFCJ/ipZi8+/r0VRd95WcvrGhNicufWNSfmzoWSJaFTJ6hVC8aOhRMnAvxk0dG63sDMmbB2rVYd/f57aNsWSpTQJqI33tD1T3fvDvCTm4zwqxqoiLQC3geigCHOub4p7s8LjABqAnuA251zm737ngW6AglAL+fcNH+OmZpgVgPduxe++ELX0Fi5UhdxefaWtTw04Xqi9+3SVd27dAlKbMZkREICjBihE33XrdOE0LmzNg/VqOHHGsOZceyY9hXExcGMGbB+/en7YmJ0kePLLkv+FRNj1RIDJK1qoGdNACISBawDrgfigcVAZ+fcKp99egDXOOceEpFOwM3OudtFpBIwCqgDXATMBC73HpbuMVOT3Qlgzx7ty5o8Wb+OHoU6dRwPNt/MnYsfJ/+MOKhYEcaP12EXxoSRhASd7DtiBHz9NRw/rp+7zZpB06Z6lVsqq8Yx7N8PS5dq0bklS3TW2saNOnY1kYg2H116qQ5DTfwqXVqzVokSp7/OOy+LMldkOJcEUB94xTnX0vv5WQDn3H989pnm7TNfRHIDfwElgT6++ybu5z0s3WOmJlAJwDkdHpeQoF8HDsCOHfq1bZsutj1/vmPBAjh1SriwxHFuqfEHD17wNdUWfqK1d4sX1zn3jzyipXSNCWN//63N9t98o8NH//lHt5cuDZdfrpOAy5fXCrYlS+rXeedBvnx6kp437+nb0dGZ/Cx2TtfN2Lgx+Vd8vP5jbt0Khw6l/tjo6OQJIbWv4sV19FJ0NOTJk/x7dLTW1s6VS4MXyfrb2Ziw0koAuf147MXAFp+f44GUQ1yS9nHOnRSR/UAJb/uCFI+92Lt9tmMGTO3a2myTkKCr351t+by8HKUqy3mB72jLN9Tc8zO5Zjj9Y2naFJ58Eu6809ZxNBGjWDGt8tCtm/6PLF2qiWDVKm0mmjBBr4j9lTt38s8536/ixbWf+AwieslRqhRpzp48cECTwe7dGlDiV8qf1607fTvgnRwB5E+ySLRrl2bZAPInAaSWplJeNqS1T1rbU+t8TvVSRES6Ad28Hw+KyNo04kzP+YDfvU3HgEXe16u+dxw/rsMpvvtO6+5mrwy9hhBlryE0ZPlrSG+Z4cOHA3LyGxnvg3O7/Z4vkT//uTzXpalt9CcBxANlfX4uA2xLY594rwmoCLD3LI892zEBcM4NAgb5EWeaRGRJapc/4cReQ2iw1xAa7DUEhj/DQBcDFUWknIjkAToBcSn2iQPu9W7fCsxy2rkQB3QSkbwiUg6oiJ5Y+3NMY4wxWeisVwBem35PYBo6ZHOYc26liLwKLHHOxQFDgc9FZAN65t/Je+xKERkLrAJOAo845xIAUjtm4F+eMcaYtPg1DyDciUg3rykpbNlrCA32GkKDvYYAxZATEoAxxpgzWSkIY4zJoSI+AYhIKxFZKyIbRKRPsOPxl4hsFpHfRGSZiCzxthUXkRkist77XizYcfoSkWEislNEVvhsSzVmUR9478uvIlIjeJGflsZreEVEtnrvxTIRae1z37Pea1grIi2DE3VyIlJWRH4QkdUislJEHvO2h817kc5rCJv3QkTyicgiEVnuvYb/87aXE5GF3vswxhsIgzdYZoz3GhaKSEyWB+mci9gvtIN5I1AeyAMsByoFOy4/Y98MnJ9i21tAH+92H+DNYMeZIr5YoAaw4mwxA62BqehckXrAwmDHn85reAV4KpV9K3l/U3mBct7fWlQIvIbSQA3vdiG07EqlcHov0nkNYfNeeL/P87zb0cBC7/c7FujkbR8IPOzd7gEM9G53AsZkdYyRfgVQB9jgnPvdOXccGA20D3JM56I98Jl3+zPgpiDGcgbn3Bx0FJivtGJuD4xwagFQVERKZ0+kaUvjNaSlPTDaOXfMObcJ2ID+zQWVc267c26pd/sAsBqdgR8270U6ryEtIfdeeL/Pg96P0d6XA5oC473tKd+HxPdnPNBMJGvrRUR6AkitjEV6f0ShxAHTReRnbzY0QCnn3HbQfxDggqBF57+0Yg6396an1zwyzKfpLeRfg9eMUB09+wzL9yLFa4Awei9EJEpElgE7gRnolck+51ziXGnfOJOV1AESS+pkmUhPAP6UsQhVDZ1zNYAbgEdEJDbYAQVYOL03A4DLgGrAduC/3vaQfg0ich4wAXjcOfdPerumsi0kXkcqryGs3gvnXIJzrhpa7aAOcFVqu3nfs/01RHoC8KeMRUhyzm3zvu8EJqF/PDsSL8297zuDF6Hf0oo5bN4b59wO7x/5FDCY000LIfsaRCQa/eAc6Zyb6G0Oq/citdcQju8FgHNuHzAb7QMoKloyB5LHmfQaJHlJnSwT6QkgLEtOiEhBESmUeBtoAawgecmNe4GvgxNhhqQVcxxwjzcCpR6wP7F5ItSkaA+/GX0vIO1SJ0HltRsPBVY7597xuSts3ou0XkM4vRciUlJEinq38wPN0b6MH9CSOXDm+5BaSZ2sE8xe8uz4Qkc4rEPb3p4Pdjx+xlweHdGwHFiZGDfaHvg9sN77XjzYsaaIexR6WX4CPZvpmlbM6OVuf+99+Q2oFez403kNn3sx/or+k5b22f957zWsBW4IdvxeTNeiTQe/Asu8r9bh9F6k8xrC5r0ArgF+8WJdAbzkbS+PJqcNwDggr7c9n/fzBu/+8lkdo80ENsaYHCrSm4CMMcakwRKAMcbkUJYAjDEmh7IEYIwxOZQlAGOMyaEsARiTSSIyXERuPfuexoQmSwDGZBOf2Z/GhAT7gzTGhzfzeiw6RT8KeA24ArgRyA/MA7q7FBNoROSl1PYRkdnezw2BWSLSBbjcOXdCRAqjk4QqOudOZMPLMyYZuwIwJrlWwDbnXFXnXGXgO+Aj51xt7+f8QNtUHpfePkWdc42dc/+H1oNp423vBEywD38TLJYAjEnuN6C5iLwpIo2cc/uB67wVmn5Da7lfncrj0ttnjM/tIcB93u37gE8D/xKM8Y81ARnjwzm3TkRqonVn/iMi04FH0Po4W0TkFbRmSxIRyQd8nM4+h3yO/5OIxIhIY3TFqhUYEyR2BWCMDxG5CDjsnPsC6IcuDwmw26tNn9qon3x+7ONrBFp0zs7+TVDZFYAxyVUB3haRRczVVgAAAHdJREFUU2hF0IfRJft+Q9dpXpzyAc65fSIyOL19UhgJvI4mAWOCxqqBGpPNvLkD7Z1zdwc7FpOz2RWAMdlIRD5El/lsHexYjLErAGOMyaGsE9gYY3IoSwDGGJNDWQIwxpgcyhKAMcbkUJYAjDEmh7IEYIwxOdT/A/k2rXh6uDk0AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot the real vs predicted\n",
    "ax1 = sns.distplot(test_y, hist=False, color=\"r\", label=\"Real Values\")\n",
    "sns.distplot(test_x_predict, hist=False, color=\"b\", label=\"Predicted Values\" , ax=ax1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">3.3. Gradient Boosting Regressor</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "gbr = GradientBoostingRegressor(n_estimators=150, max_depth=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingRegressor(alpha=0.9, ccp_alpha=0.0, criterion='friedman_mse',\n",
       "                          init=None, learning_rate=0.1, loss='ls', max_depth=5,\n",
       "                          max_features=None, max_leaf_nodes=None,\n",
       "                          min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                          min_samples_leaf=1, min_samples_split=2,\n",
       "                          min_weight_fraction_leaf=0.0, n_estimators=150,\n",
       "                          n_iter_no_change=None, presort='deprecated',\n",
       "                          random_state=None, subsample=1.0, tol=0.0001,\n",
       "                          validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#fit the model\n",
    "gbr.fit(train_x,train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GradientBoostingRegressor(alpha=0.9, ccp_alpha=0.0, criterion='friedman_mse',\n",
      "                          init=None, learning_rate=0.1, loss='ls', max_depth=5,\n",
      "                          max_features=None, max_leaf_nodes=None,\n",
      "                          min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                          min_samples_leaf=1, min_samples_split=2,\n",
      "                          min_weight_fraction_leaf=0.0, n_estimators=150,\n",
      "                          n_iter_no_change=None, presort='deprecated',\n",
      "                          random_state=None, subsample=1.0, tol=0.0001,\n",
      "                          validation_fraction=0.1, verbose=0, warm_start=False)\n",
      "Mean: 357.565913933955, Std: 0.32566049342847236\n"
     ]
    }
   ],
   "source": [
    "#evaluate the model\n",
    "cv(gbr,train_x,train_y,2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "#calculate the prediction\n",
    "test_x_gbrpredict = gbr.predict(test_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x62006fa288>"
      ]
     },
     "execution_count": 28,
     "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": [
    "#plot the real vs predicted\n",
    "ax1 = sns.distplot(test_y, hist=False, color=\"r\", label=\"Real Values\")\n",
    "sns.distplot(test_x_gbrpredict, hist=False, color=\"b\", label=\"Predicted Values\" , ax=ax1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### <span style=\"color:black\">3.1.1 Observations</span>  \n",
    "> Although, Predicted Vs Real plots for both models almost looks same but based on MSE, Gradient Boosting Regressor did better than Linear Regression.  \n",
    "> Although GBR can be tuneed further by increasing Number of Estimators parameter (n_estimators) and Maximum Depth parameter (max_depth) but that would at expense of performance.   \n",
    "    > (Number of Estimators: It is denoted as n_estimators. The default value of this parameter is 100.Number of estimators is basically the number of boosting stages to be performed by the model. In other words number of estimators denotes the number of trees in the forest. More number of trees helps in learning the data better. On the other hand, more number of trees can result in higher training time. Hence we need to find the right and balanced value of n_estimators for optimal performance.)  \n",
    "    > (Maximum Depth: It is denoted as max_depth. The default value of max_depth is 3 and it is an optional parameter. The maximum depth is the depth of the decision tree estimator in the gradient boosting regressor. We need to find the optimum value of this hyperparameter for best performance. As an example the best value of this parameter may depend on the input variables.)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">3.3.1 feature importance</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Feature importance')"
      ]
     },
     "execution_count": 29,
     "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": [
    "#visualization of feature importance using GBR\n",
    "fi = pd.Series(gbr.feature_importances_, index=train_merge_Hot_Enc.columns).nlargest(10).plot(kind='bar')\n",
    "plt.xlabel('Feature importance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### <span style=\"color:black\">3.3.1.1 Observations</span> \n",
    "> Plot shows Job type and Years of Experience as the important features effecting the model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## <span style=\"color:black\">4. Deploy</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">4.1. transform the features data for prediction</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "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>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "      <th>jobType_CEO</th>\n",
       "      <th>jobType_CFO</th>\n",
       "      <th>jobType_CTO</th>\n",
       "      <th>jobType_JANITOR</th>\n",
       "      <th>jobType_JUNIOR</th>\n",
       "      <th>jobType_MANAGER</th>\n",
       "      <th>jobType_SENIOR</th>\n",
       "      <th>jobType_VICE_PRESIDENT</th>\n",
       "      <th>...</th>\n",
       "      <th>major_MATH</th>\n",
       "      <th>major_NONE</th>\n",
       "      <th>major_PHYSICS</th>\n",
       "      <th>industry_AUTO</th>\n",
       "      <th>industry_EDUCATION</th>\n",
       "      <th>industry_FINANCE</th>\n",
       "      <th>industry_HEALTH</th>\n",
       "      <th>industry_OIL</th>\n",
       "      <th>industry_SERVICE</th>\n",
       "      <th>industry_WEB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>22</td>\n",
       "      <td>73</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20</td>\n",
       "      <td>47</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>17</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>14</td>\n",
       "      <td>96</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>44</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   yearsExperience  milesFromMetropolis  jobType_CEO  jobType_CFO  \\\n",
       "0               22                   73            0            0   \n",
       "1               20                   47            0            0   \n",
       "2               17                    9            0            0   \n",
       "3               14                   96            0            0   \n",
       "4               10                   44            0            0   \n",
       "\n",
       "   jobType_CTO  jobType_JANITOR  jobType_JUNIOR  jobType_MANAGER  \\\n",
       "0            0                0               0                1   \n",
       "1            0                0               1                0   \n",
       "2            1                0               0                0   \n",
       "3            0                0               0                1   \n",
       "4            0                0               1                0   \n",
       "\n",
       "   jobType_SENIOR  jobType_VICE_PRESIDENT  ...  major_MATH  major_NONE  \\\n",
       "0               0                       0  ...           0           1   \n",
       "1               0                       0  ...           0           1   \n",
       "2               0                       0  ...           0           0   \n",
       "3               0                       0  ...           0           1   \n",
       "4               0                       0  ...           0           0   \n",
       "\n",
       "   major_PHYSICS  industry_AUTO  industry_EDUCATION  industry_FINANCE  \\\n",
       "0              0              0                   0                 0   \n",
       "1              0              1                   0                 0   \n",
       "2              0              0                   0                 0   \n",
       "3              0              0                   0                 0   \n",
       "4              0              0                   0                 0   \n",
       "\n",
       "   industry_HEALTH  industry_OIL  industry_SERVICE  industry_WEB  \n",
       "0                1             0                 0             0  \n",
       "1                0             0                 0             0  \n",
       "2                1             0                 0             0  \n",
       "3                0             1                 0             0  \n",
       "4                0             1                 0             0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#massage the test data\n",
    "test_features_df.drop(['companyId', 'jobId'],axis = 1)\n",
    "test_features_df_Hot_Enc = test_features_df[num_features].join(pd.get_dummies(test_features_df[cat_features]))\n",
    "test_features_df_Hot_Enc.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([112.25194555,  93.2195217 , 181.49761349, ...,  54.50038396,\n",
       "       160.62274084, 113.72743758])"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "##calculate the prediction\n",
    "test_features_df_Hot_Enc_gbrpredict = gbr.predict(test_features_df_Hot_Enc)\n",
    "test_features_df_Hot_Enc_gbrpredict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "#convert the array to dataframe\n",
    "test_features_df_Hot_Enc_gbrpredict_df = pd.DataFrame(data=test_features_df_Hot_Enc_gbrpredict,  columns=[\"Predicted_salary\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### <span style=\"color:black\">4.2. Final</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "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>jobId</th>\n",
       "      <th>companyId</th>\n",
       "      <th>jobType</th>\n",
       "      <th>degree</th>\n",
       "      <th>major</th>\n",
       "      <th>industry</th>\n",
       "      <th>yearsExperience</th>\n",
       "      <th>milesFromMetropolis</th>\n",
       "      <th>Predicted_salary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>JOB1362685407687</td>\n",
       "      <td>COMP33</td>\n",
       "      <td>MANAGER</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>22</td>\n",
       "      <td>73</td>\n",
       "      <td>112.251946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>JOB1362685407688</td>\n",
       "      <td>COMP13</td>\n",
       "      <td>JUNIOR</td>\n",
       "      <td>NONE</td>\n",
       "      <td>NONE</td>\n",
       "      <td>AUTO</td>\n",
       "      <td>20</td>\n",
       "      <td>47</td>\n",
       "      <td>93.219522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>JOB1362685407689</td>\n",
       "      <td>COMP10</td>\n",
       "      <td>CTO</td>\n",
       "      <td>MASTERS</td>\n",
       "      <td>BIOLOGY</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>17</td>\n",
       "      <td>9</td>\n",
       "      <td>181.497613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>JOB1362685407690</td>\n",
       "      <td>COMP21</td>\n",
       "      <td>MANAGER</td>\n",
       "      <td>HIGH_SCHOOL</td>\n",
       "      <td>NONE</td>\n",
       "      <td>OIL</td>\n",
       "      <td>14</td>\n",
       "      <td>96</td>\n",
       "      <td>104.041410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>JOB1362685407691</td>\n",
       "      <td>COMP36</td>\n",
       "      <td>JUNIOR</td>\n",
       "      <td>DOCTORAL</td>\n",
       "      <td>BIOLOGY</td>\n",
       "      <td>OIL</td>\n",
       "      <td>10</td>\n",
       "      <td>44</td>\n",
       "      <td>115.046296</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>999995</th>\n",
       "      <td>JOB1362686407682</td>\n",
       "      <td>COMP54</td>\n",
       "      <td>VICE_PRESIDENT</td>\n",
       "      <td>BACHELORS</td>\n",
       "      <td>MATH</td>\n",
       "      <td>OIL</td>\n",
       "      <td>14</td>\n",
       "      <td>3</td>\n",
       "      <td>166.651740</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999996</th>\n",
       "      <td>JOB1362686407683</td>\n",
       "      <td>COMP5</td>\n",
       "      <td>MANAGER</td>\n",
       "      <td>NONE</td>\n",
       "      <td>NONE</td>\n",
       "      <td>HEALTH</td>\n",
       "      <td>20</td>\n",
       "      <td>67</td>\n",
       "      <td>106.297231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999997</th>\n",
       "      <td>JOB1362686407684</td>\n",
       "      <td>COMP61</td>\n",
       "      <td>JANITOR</td>\n",
       "      <td>NONE</td>\n",
       "      <td>NONE</td>\n",
       "      <td>OIL</td>\n",
       "      <td>1</td>\n",
       "      <td>91</td>\n",
       "      <td>54.500384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999998</th>\n",
       "      <td>JOB1362686407685</td>\n",
       "      <td>COMP19</td>\n",
       "      <td>CTO</td>\n",
       "      <td>DOCTORAL</td>\n",
       "      <td>MATH</td>\n",
       "      <td>OIL</td>\n",
       "      <td>14</td>\n",
       "      <td>63</td>\n",
       "      <td>160.622741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999999</th>\n",
       "      <td>JOB1362686407686</td>\n",
       "      <td>COMP34</td>\n",
       "      <td>JUNIOR</td>\n",
       "      <td>NONE</td>\n",
       "      <td>NONE</td>\n",
       "      <td>OIL</td>\n",
       "      <td>16</td>\n",
       "      <td>31</td>\n",
       "      <td>113.727438</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1000000 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                   jobId companyId         jobType       degree    major  \\\n",
       "0       JOB1362685407687    COMP33         MANAGER  HIGH_SCHOOL     NONE   \n",
       "1       JOB1362685407688    COMP13          JUNIOR         NONE     NONE   \n",
       "2       JOB1362685407689    COMP10             CTO      MASTERS  BIOLOGY   \n",
       "3       JOB1362685407690    COMP21         MANAGER  HIGH_SCHOOL     NONE   \n",
       "4       JOB1362685407691    COMP36          JUNIOR     DOCTORAL  BIOLOGY   \n",
       "...                  ...       ...             ...          ...      ...   \n",
       "999995  JOB1362686407682    COMP54  VICE_PRESIDENT    BACHELORS     MATH   \n",
       "999996  JOB1362686407683     COMP5         MANAGER         NONE     NONE   \n",
       "999997  JOB1362686407684    COMP61         JANITOR         NONE     NONE   \n",
       "999998  JOB1362686407685    COMP19             CTO     DOCTORAL     MATH   \n",
       "999999  JOB1362686407686    COMP34          JUNIOR         NONE     NONE   \n",
       "\n",
       "       industry  yearsExperience  milesFromMetropolis  Predicted_salary  \n",
       "0        HEALTH               22                   73        112.251946  \n",
       "1          AUTO               20                   47         93.219522  \n",
       "2        HEALTH               17                    9        181.497613  \n",
       "3           OIL               14                   96        104.041410  \n",
       "4           OIL               10                   44        115.046296  \n",
       "...         ...              ...                  ...               ...  \n",
       "999995      OIL               14                    3        166.651740  \n",
       "999996   HEALTH               20                   67        106.297231  \n",
       "999997      OIL                1                   91         54.500384  \n",
       "999998      OIL               14                   63        160.622741  \n",
       "999999      OIL               16                   31        113.727438  \n",
       "\n",
       "[1000000 rows x 9 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#merge predicted dataframe with features data frame.\n",
    "test_features_target_df = test_features_df.join(test_features_df_Hot_Enc_gbrpredict_df)\n",
    "test_features_target_df"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}