dp_test.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import pandas as pd\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from tqdm.notebook import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>PERIODO</th>\n",
       "      <th>ESTU_PRGM_ACADEMICO</th>\n",
       "      <th>ESTU_PRGM_DEPARTAMENTO</th>\n",
       "      <th>ESTU_VALORMATRICULAUNIVERSIDAD_Entre 2.5 millones y menos de 4 millones</th>\n",
       "      <th>ESTU_VALORMATRICULAUNIVERSIDAD_Entre 1 millón y menos de 2.5 millones</th>\n",
       "      <th>ESTU_VALORMATRICULAUNIVERSIDAD_Más de 7 millones</th>\n",
       "      <th>ESTU_VALORMATRICULAUNIVERSIDAD_Menos de 500 mil</th>\n",
       "      <th>ESTU_VALORMATRICULAUNIVERSIDAD_Entre 500 mil y menos de 1 millón</th>\n",
       "      <th>ESTU_VALORMATRICULAUNIVERSIDAD_Entre 4 millones y menos de 5.5 millones</th>\n",
       "      <th>ESTU_VALORMATRICULAUNIVERSIDAD_Entre 5.5 millones y menos de 7 millones</th>\n",
       "      <th>...</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_Ninguno</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_Primaria incompleta</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_Educación profesional completa</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_Secundaria (Bachillerato) incompleta</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_Educación profesional incompleta</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_Técnica o tecnológica incompleta</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_No Aplica</th>\n",
       "      <th>FAMI_EDUCACIONMADRE_No sabe</th>\n",
       "      <th>ESTU_PAGOMATRICULAPROPIO</th>\n",
       "      <th>RENDIMIENTO_GLOBAL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20183.0</td>\n",
       "      <td>1.664298</td>\n",
       "      <td>1.533911</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20195.0</td>\n",
       "      <td>1.791667</td>\n",
       "      <td>1.460920</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>20195.0</td>\n",
       "      <td>1.640811</td>\n",
       "      <td>1.426440</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>20195.0</td>\n",
       "      <td>1.447136</td>\n",
       "      <td>1.406953</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>20183.0</td>\n",
       "      <td>0.396624</td>\n",
       "      <td>1.460920</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   PERIODO  ESTU_PRGM_ACADEMICO  ESTU_PRGM_DEPARTAMENTO  \\\n",
       "0  20183.0             1.664298                1.533911   \n",
       "1  20195.0             1.791667                1.460920   \n",
       "2  20195.0             1.640811                1.426440   \n",
       "3  20195.0             1.447136                1.406953   \n",
       "4  20183.0             0.396624                1.460920   \n",
       "\n",
       "   ESTU_VALORMATRICULAUNIVERSIDAD_Entre 2.5 millones y menos de 4 millones  \\\n",
       "0                                                1.0                         \n",
       "1                                                0.0                         \n",
       "2                                                1.0                         \n",
       "3                                                0.0                         \n",
       "4                                                0.0                         \n",
       "\n",
       "   ESTU_VALORMATRICULAUNIVERSIDAD_Entre 1 millón y menos de 2.5 millones  \\\n",
       "0                                                0.0                       \n",
       "1                                                1.0                       \n",
       "2                                                0.0                       \n",
       "3                                                1.0                       \n",
       "4                                                0.0                       \n",
       "\n",
       "   ESTU_VALORMATRICULAUNIVERSIDAD_Más de 7 millones  \\\n",
       "0                                               0.0   \n",
       "1                                               0.0   \n",
       "2                                               0.0   \n",
       "3                                               0.0   \n",
       "4                                               1.0   \n",
       "\n",
       "   ESTU_VALORMATRICULAUNIVERSIDAD_Menos de 500 mil  \\\n",
       "0                                              0.0   \n",
       "1                                              0.0   \n",
       "2                                              0.0   \n",
       "3                                              0.0   \n",
       "4                                              0.0   \n",
       "\n",
       "   ESTU_VALORMATRICULAUNIVERSIDAD_Entre 500 mil y menos de 1 millón  \\\n",
       "0                                                0.0                  \n",
       "1                                                0.0                  \n",
       "2                                                0.0                  \n",
       "3                                                0.0                  \n",
       "4                                                0.0                  \n",
       "\n",
       "   ESTU_VALORMATRICULAUNIVERSIDAD_Entre 4 millones y menos de 5.5 millones  \\\n",
       "0                                                0.0                         \n",
       "1                                                0.0                         \n",
       "2                                                0.0                         \n",
       "3                                                0.0                         \n",
       "4                                                0.0                         \n",
       "\n",
       "   ESTU_VALORMATRICULAUNIVERSIDAD_Entre 5.5 millones y menos de 7 millones  \\\n",
       "0                                                0.0                         \n",
       "1                                                0.0                         \n",
       "2                                                0.0                         \n",
       "3                                                0.0                         \n",
       "4                                                0.0                         \n",
       "\n",
       "   ...  FAMI_EDUCACIONMADRE_Ninguno  FAMI_EDUCACIONMADRE_Primaria incompleta  \\\n",
       "0  ...                          0.0                                      0.0   \n",
       "1  ...                          0.0                                      0.0   \n",
       "2  ...                          0.0                                      0.0   \n",
       "3  ...                          0.0                                      0.0   \n",
       "4  ...                          0.0                                      0.0   \n",
       "\n",
       "   FAMI_EDUCACIONMADRE_Educación profesional completa  \\\n",
       "0                                                0.0    \n",
       "1                                                0.0    \n",
       "2                                                0.0    \n",
       "3                                                0.0    \n",
       "4                                                0.0    \n",
       "\n",
       "   FAMI_EDUCACIONMADRE_Secundaria (Bachillerato) incompleta  \\\n",
       "0                                                0.0          \n",
       "1                                                0.0          \n",
       "2                                                0.0          \n",
       "3                                                0.0          \n",
       "4                                                0.0          \n",
       "\n",
       "   FAMI_EDUCACIONMADRE_Educación profesional incompleta  \\\n",
       "0                                                0.0      \n",
       "1                                                0.0      \n",
       "2                                                0.0      \n",
       "3                                                0.0      \n",
       "4                                                0.0      \n",
       "\n",
       "   FAMI_EDUCACIONMADRE_Técnica o tecnológica incompleta  \\\n",
       "0                                                0.0      \n",
       "1                                                0.0      \n",
       "2                                                0.0      \n",
       "3                                                0.0      \n",
       "4                                                0.0      \n",
       "\n",
       "   FAMI_EDUCACIONMADRE_No Aplica  FAMI_EDUCACIONMADRE_No sabe  \\\n",
       "0                            0.0                          0.0   \n",
       "1                            0.0                          0.0   \n",
       "2                            0.0                          0.0   \n",
       "3                            0.0                          0.0   \n",
       "4                            0.0                          0.0   \n",
       "\n",
       "   ESTU_PAGOMATRICULAPROPIO  RENDIMIENTO_GLOBAL  \n",
       "0                       0.0                   1  \n",
       "1                       0.0                   1  \n",
       "2                       1.0                   0  \n",
       "3                       1.0                   2  \n",
       "4                       0.0                   0  \n",
       "\n",
       "[5 rows x 50 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = pd.read_csv('data/dataset_transformed.csv')\n",
    "dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  1%|          | 1/100 [00:12<19:54, 12.07s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train| Loss: 2.8907705887360606, acc: 0.25110421478060047\n",
      "Test| Loss: 0.002568064140657657, acc: 0.0\n",
      "--------------------------------------------------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  2%|▏         | 2/100 [00:23<19:02, 11.66s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train| Loss: 1.3866546190363156, acc: 0.2505972142032333\n",
      "Test| Loss: 0.0025506439684061986, acc: 0.00046125461254612545\n",
      "--------------------------------------------------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  3%|▎         | 3/100 [00:36<19:30, 12.07s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train| Loss: 1.3867040440336937, acc: 0.250196665704388\n",
      "Test| Loss: 0.0025886501773257097, acc: 0.0\n",
      "--------------------------------------------------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  4%|▍         | 4/100 [00:48<19:32, 12.21s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train| Loss: 1.38663926240203, acc: 0.25051963048498843\n",
      "Test| Loss: 0.0025841303417163584, acc: 0.00046125461254612545\n",
      "--------------------------------------------------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  5%|▌         | 5/100 [01:00<19:15, 12.16s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train| Loss: 1.3866433897414878, acc: 0.2506008227482679\n",
      "Test| Loss: 0.0025566268671042805, acc: 0.0009225092250922509\n",
      "--------------------------------------------------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  6%|▌         | 6/100 [01:11<18:28, 11.79s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train| Loss: 1.3866711177671753, acc: 0.25117458140877597\n",
      "Test| Loss: 0.002559631952940318, acc: 0.00046125461254612545\n",
      "--------------------------------------------------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  6%|▌         | 6/100 [01:23<21:51, 13.95s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Early stopping at epoch 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "import torch.utils\n",
    "from torch.utils.data import DataLoader, TensorDataset\n",
    "from torch.optim.lr_scheduler import ReduceLROnPlateau\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "x = dataset.drop('RENDIMIENTO_GLOBAL', axis=1).values\n",
    "y = dataset['RENDIMIENTO_GLOBAL'].values\n",
    "\n",
    "x = torch.tensor(x, dtype=torch.float32)\n",
    "y = torch.tensor(y, dtype=torch.float32)\n",
    "\n",
    "train_size = int(0.8 * len(x))\n",
    "test_size = len(x) - train_size\n",
    "train_dataset, test_dataset = torch.utils.data.random_split(TensorDataset(x, y), [train_size, test_size], generator=torch.Generator().manual_seed(23))\n",
    "\n",
    "batch_size = 256\n",
    "\n",
    "train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)\n",
    "test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False)\n",
    "\n",
    "\n",
    "\n",
    "class NNModel(nn.Module):\n",
    "    def __init__(self,input_layer, output_layer, hidden_layers = 20, *args, **kwargs):\n",
    "        super().__init__(*args, **kwargs)\n",
    "        self.sequiential_layer = nn.Sequential(\n",
    "            nn.Linear(input_layer,hidden_layers),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(hidden_layers,hidden_layers//2),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(hidden_layers//2,output_layer),\n",
    "            nn.Dropout(0.5)\n",
    "        )\n",
    "        self.sequiential_layer2 = nn.Sequential(\n",
    "            nn.Linear(hidden_layers,hidden_layers//2),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(hidden_layers//2,hidden_layers//2),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(hidden_layers//2,hidden_layers//4),\n",
    "            nn.Dropout(0.5)\n",
    "        )\n",
    "        self.sequiential_layer3 = nn.Sequential(\n",
    "            nn.Linear(hidden_layers//4,hidden_layers//8),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(hidden_layers//8,hidden_layers//8),\n",
    "            nn.ReLU(),\n",
    "            nn.Dropout(0.5),\n",
    "            nn.Linear(hidden_layers//8,output_layer)\n",
    "        )\n",
    "    def forward(self, x:torch.Tensor):\n",
    "        x = self.sequiential_layer(x)\n",
    "        #x = self.sequiential_layer2(x)\n",
    "        #x = self.sequiential_layer3(x)\n",
    "        return x\n",
    "\n",
    "\n",
    "input_size = 49\n",
    "\n",
    "\n",
    "torch.manual_seed(23)\n",
    "num_classes = len(dataset['RENDIMIENTO_GLOBAL'].unique())\n",
    "\n",
    "model = NNModel(49, num_classes, 100).to(device)\n",
    "\n",
    "optimizer = optim.AdamW(model.parameters(), lr=1e-2, weight_decay=1e-5)\n",
    "#scheduler = ReduceLROnPlateau(optimizer, 'min', patience=5, factor=0.5, verbose=True)\n",
    "loss_fn = nn.CrossEntropyLoss()\n",
    "\n",
    "\n",
    "num_epochs = 100\n",
    "\n",
    "\n",
    "patience = 5\n",
    "best_loss = float('inf')\n",
    "counter = 0\n",
    "\n",
    "for epoch in tqdm(range(num_epochs)):\n",
    "\n",
    "    train_loss, train_acc = 0, 0\n",
    "\n",
    "    model.train()\n",
    "    for batch, (x, y) in enumerate(train_loader):\n",
    "        x = x.to(device)\n",
    "        y = y.to(device).long()\n",
    "        y_logit = model(x)\n",
    "        loss = loss_fn(y_logit, y)\n",
    "        train_loss += loss.item()\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        y_pred = torch.softmax(y_logit, dim=1).argmax(dim=1)\n",
    "        train_acc += accuracy_score(y.cpu().numpy(), y_pred.cpu().numpy())\n",
    "\n",
    "    train_loss /= len(train_loader)\n",
    "    train_acc /= len(train_loader)\n",
    "\n",
    "    model.eval()\n",
    "    with torch.inference_mode():\n",
    "        for batch, (x, y) in enumerate(test_loader):\n",
    "\n",
    "            test_loss, test_acc = 0, 0\n",
    "            x = x.to(device)\n",
    "            y = y.to(device).long()\n",
    "            y_logit = model(x)\n",
    "            loss = loss_fn(y_logit, y)\n",
    "            test_loss += loss.item()\n",
    "            y_pred = torch.softmax(y_logit, dim=1).argmax(dim=1)\n",
    "            test_acc += accuracy_score(y.cpu().numpy(), y_pred.cpu().numpy())\n",
    "\n",
    "        test_loss /= len(test_loader)\n",
    "        test_acc /= len(test_loader)\n",
    "    \n",
    "    #scheduler.step(test_loss)\n",
    "\n",
    "    if test_loss < best_loss:\n",
    "        best_loss = test_loss\n",
    "        counter = 0\n",
    "    else:\n",
    "        counter += 1\n",
    "        if counter >= patience:\n",
    "            print(f'Early stopping at epoch {epoch}')\n",
    "            break\n",
    "    \n",
    "    print(f'Train| Loss: {train_loss}, acc: {train_acc}')\n",
    "    print(f'Test| Loss: {test_loss}, acc: {test_acc}')\n",
    "\n",
    "    print(\"-\"*50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "model.eval()\n",
    "all_preds = []\n",
    "all_labels = []\n",
    "with torch.no_grad():\n",
    "    for inputs, labels in test_loader:\n",
    "        outputs = model(inputs.to(device))\n",
    "        predicted = torch.softmax(outputs, dim=1).argmax(dim=1)\n",
    "        all_preds.extend(predicted.cpu().numpy())\n",
    "        all_labels.extend(labels.numpy())\n",
    "\n",
    "# Paso 2: Calcular la matriz de confusión\n",
    "cm = confusion_matrix(all_labels, all_preds)\n",
    "\n",
    "# Paso 3: Visualizar la matriz de confusión\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm)\n",
    "disp.plot(cmap=plt.cm.Blues)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pytorch",
   "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": "3.12.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}