Classification_CustomNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Basic ML Solutions in PyTorch\n",
    "# Classification Algorithim using Custom NN\n",
    "#### Titanic Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import torch"
   ]
  },
  {
   "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>PassengerId</th>\n",
       "      <th>Survived</th>\n",
       "      <th>Pclass</th>\n",
       "      <th>Name</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Ticket</th>\n",
       "      <th>Fare</th>\n",
       "      <th>Cabin</th>\n",
       "      <th>Embarked</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>177</th>\n",
       "      <td>178</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Isham, Miss. Ann Elizabeth</td>\n",
       "      <td>female</td>\n",
       "      <td>50.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>PC 17595</td>\n",
       "      <td>28.7125</td>\n",
       "      <td>C49</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>370</th>\n",
       "      <td>371</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Harder, Mr. George Achilles</td>\n",
       "      <td>male</td>\n",
       "      <td>25.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11765</td>\n",
       "      <td>55.4417</td>\n",
       "      <td>E50</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>505</th>\n",
       "      <td>506</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Penasco y Castellana, Mr. Victor de Satode</td>\n",
       "      <td>male</td>\n",
       "      <td>18.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>PC 17758</td>\n",
       "      <td>108.9000</td>\n",
       "      <td>C65</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>Vestrom, Miss. Hulda Amanda Adolfina</td>\n",
       "      <td>female</td>\n",
       "      <td>14.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>350406</td>\n",
       "      <td>7.8542</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>194</th>\n",
       "      <td>195</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Brown, Mrs. James Joseph (Margaret Tobin)</td>\n",
       "      <td>female</td>\n",
       "      <td>44.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>PC 17610</td>\n",
       "      <td>27.7208</td>\n",
       "      <td>B4</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     PassengerId  Survived  Pclass  \\\n",
       "177          178         0       1   \n",
       "370          371         1       1   \n",
       "505          506         0       1   \n",
       "14            15         0       3   \n",
       "194          195         1       1   \n",
       "\n",
       "                                           Name     Sex   Age  SibSp  Parch  \\\n",
       "177                  Isham, Miss. Ann Elizabeth  female  50.0      0      0   \n",
       "370                 Harder, Mr. George Achilles    male  25.0      1      0   \n",
       "505  Penasco y Castellana, Mr. Victor de Satode    male  18.0      1      0   \n",
       "14         Vestrom, Miss. Hulda Amanda Adolfina  female  14.0      0      0   \n",
       "194   Brown, Mrs. James Joseph (Margaret Tobin)  female  44.0      0      0   \n",
       "\n",
       "       Ticket      Fare Cabin Embarked  \n",
       "177  PC 17595   28.7125   C49        C  \n",
       "370     11765   55.4417   E50        C  \n",
       "505  PC 17758  108.9000   C65        C  \n",
       "14     350406    7.8542   NaN        S  \n",
       "194  PC 17610   27.7208    B4        C  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "titanic = pd.read_csv('titanic_data/train.csv')\n",
    "titanic.sample(5)"
   ]
  },
  {
   "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>Survived</th>\n",
       "      <th>Pclass</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Age</th>\n",
       "      <th>Fare</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>731</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>male</td>\n",
       "      <td>11.0</td>\n",
       "      <td>18.7875</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>395</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>male</td>\n",
       "      <td>22.0</td>\n",
       "      <td>7.7958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>female</td>\n",
       "      <td>17.0</td>\n",
       "      <td>10.5000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>789</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>male</td>\n",
       "      <td>46.0</td>\n",
       "      <td>79.2000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>201</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>male</td>\n",
       "      <td>NaN</td>\n",
       "      <td>69.5500</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Survived  Pclass     Sex   Age     Fare\n",
       "731         0       3    male  11.0  18.7875\n",
       "395         0       3    male  22.0   7.7958\n",
       "84          1       2  female  17.0  10.5000\n",
       "789         0       1    male  46.0  79.2000\n",
       "201         0       3    male   NaN  69.5500"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "titanic = titanic[['Survived', 'Pclass', 'Sex', 'Age', 'Fare']]\n",
    "titanic.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "titanic = titanic.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "le = LabelEncoder()\n",
    "\n",
    "titanic['Sex'] = le.fit_transform(titanic['Sex'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = titanic.drop('Survived', axis=1)\n",
    "y = titanic[['Survived']]\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)\n",
    "\n",
    "X_train = torch.from_numpy(x_train.values).float()\n",
    "X_test = torch.from_numpy(x_test.values).float()\n",
    "\n",
    "Y_train = torch.from_numpy(y_train.values).view(1, -1)[0]\n",
    "Y_test = torch.from_numpy(y_test.values).view(1, -1)[0]\n",
    "# converting (size, 1) tensor to (size) tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "inp = X_train.shape[1]\n",
    "out = 2\n",
    "hidd = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Build our own custom NN module(Net), by subclassing the nn.Module class\n",
    "\n",
    "class Net(nn.Module):\n",
    "    # within init func, we call the super class - allows us to initialize our NN before adding in layers\n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        # super keyword allows the derived class to access the parent class’s init() property\n",
    "        \n",
    "        # instantiate 3 fully connected layers\n",
    "        self.fc1 = nn.Linear(inp, hidd)\n",
    "        self.fc2 = nn.Linear(hidd, hidd)\n",
    "        self.fc3 = nn.Linear(hidd, out)\n",
    "        \n",
    "        # need to override the forward function of base nn.module class\n",
    "    def forward(self, x):\n",
    "        x = F.sigmoid(self.fc1(x))\n",
    "        x = F.sigmoid(self.fc2(x))\n",
    "        x = self.fc3(x)\n",
    "            \n",
    "        return F.log_softmax(x, dim=-1) # dim=-1 merges N-D tensors to single 1-D tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# let's instantiate this network by calling the Net class\n",
    "model = Net()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# optimizer will update our model's parameters which is why it needs it as an input argument\n",
    "\n",
    "optimizer = torch.optim.Adam(model.parameters())\n",
    "\n",
    "# Set up the loss function as NLLLoss\n",
    "\n",
    "loss_fn = nn.NLLLoss()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\PK068160\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\torch\\nn\\functional.py:1569: UserWarning: nn.functional.sigmoid is deprecated. Use torch.sigmoid instead.\n",
      "  warnings.warn(\"nn.functional.sigmoid is deprecated. Use torch.sigmoid instead.\")\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch - 100 (9%) train loss - 0.65 test loss - 0.66 accuracy - 0.5594\n",
      "Epoch - 200 (19%) train loss - 0.61 test loss - 0.61 accuracy - 0.6783\n",
      "Epoch - 300 (29%) train loss - 0.59 test loss - 0.59 accuracy - 0.6783\n",
      "Epoch - 400 (39%) train loss - 0.57 test loss - 0.56 accuracy - 0.6853\n",
      "Epoch - 500 (49%) train loss - 0.54 test loss - 0.54 accuracy - 0.7133\n",
      "Epoch - 600 (59%) train loss - 0.51 test loss - 0.50 accuracy - 0.7832\n",
      "Epoch - 700 (69%) train loss - 0.48 test loss - 0.47 accuracy - 0.8042\n",
      "Epoch - 800 (79%) train loss - 0.46 test loss - 0.45 accuracy - 0.7972\n",
      "Epoch - 900 (89%) train loss - 0.44 test loss - 0.43 accuracy - 0.7972\n",
      "Epoch - 1000 (99%) train loss - 0.43 test loss - 0.43 accuracy - 0.7972\n"
     ]
    }
   ],
   "source": [
    "# to store each epoch related data\n",
    "epoch_data = []\n",
    "epochs = 1001\n",
    "\n",
    "for epoch in range(1, epochs):\n",
    "    \n",
    "    optimizer.zero_grad()\n",
    "    \n",
    "    # configure model to training phase\n",
    "    model.train()\n",
    "    \n",
    "    Y_pred = model(X_train)\n",
    "    loss = loss_fn(Y_pred, Y_train)\n",
    "    \n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    \n",
    "    # For every epoch, calculate the loss on test data as well to check how model performs on test data\n",
    "    \n",
    "    # configure model to evaluation phase\n",
    "    model.eval()\n",
    "    \n",
    "    y_pred_test = model(X_test)\n",
    "    loss_test = loss_fn(y_pred_test, Y_test)\n",
    "    \n",
    "    # find the value with the highest probability -> our predicted value\n",
    "    _, pred = y_pred_test.data.max(1)\n",
    "    \n",
    "    # eq -> equal\n",
    "    accuracy = pred.eq(Y_test.data).sum().item() / y_test.values.size\n",
    "    \n",
    "    # store all these data in epoch_data list\n",
    "    \n",
    "    # item -> converts 1 element tensors to scalars\n",
    "    epoch_data.append([epoch, loss.data.item(), loss_test.data.item(), accuracy])\n",
    "    \n",
    "    if(epoch % 100 == 0):\n",
    "        print('Epoch - %d (%d%%) train loss - %.2f test loss - %.2f accuracy - %.4f' %(\n",
    "            epoch, epoch/epochs*100, loss.data.item(), loss_test.data.item(), accuracy))      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>test_loss</th>\n",
       "      <th>accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.735060</td>\n",
       "      <td>0.722472</td>\n",
       "      <td>0.440559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>0.732752</td>\n",
       "      <td>0.720788</td>\n",
       "      <td>0.440559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>0.730488</td>\n",
       "      <td>0.719148</td>\n",
       "      <td>0.440559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0.728266</td>\n",
       "      <td>0.717553</td>\n",
       "      <td>0.440559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>0.726091</td>\n",
       "      <td>0.716003</td>\n",
       "      <td>0.440559</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   epoch  train_loss  test_loss  accuracy\n",
       "0      1    0.735060   0.722472  0.440559\n",
       "1      2    0.732752   0.720788  0.440559\n",
       "2      3    0.730488   0.719148  0.440559\n",
       "3      4    0.728266   0.717553  0.440559\n",
       "4      5    0.726091   0.716003  0.440559"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "epoch_df = pd.DataFrame(epoch_data, columns=['epoch', 'train_loss', 'test_loss', 'accuracy'])\n",
    "epoch_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "f, (ax1, ax2) = plt.subplots(1, 2, figsize=(15,8))\n",
    "epoch_df[['train_loss', 'test_loss']].plot(ax=ax1)\n",
    "epoch_df[['accuracy']].plot(ax=ax2)\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}