part3.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PART 3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Notebook Imports\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, ConfusionMatrixDisplay\n",
    "from sklearn.experimental import enable_halving_search_cv\n",
    "from sklearn.model_selection import GridSearchCV, HalvingGridSearchCV\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "import pprint\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATASET_PATH = \"dataset.csv\"\n",
    "RANDOM_STATE = 1\n",
    "CROSS_VALIDATION_SETS = 5\n",
    "LABEL_NAMES = [\"English\", \"Urdu\", \"Mixed\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot(x, y, xlabel, ylabel, title):\n",
    "    plt.plot(x,y)\n",
    "    plt.xlabel(xlabel)\n",
    "    plt.ylabel(ylabel)\n",
    "    plt.title(title)\n",
    "    plt.xticks(x)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Importing and Splitting Data"
   ]
  },
  {
   "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>Unnamed: 0</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>LABEL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>14.443077</td>\n",
       "      <td>-8.036094</td>\n",
       "      <td>-4.077253</td>\n",
       "      <td>-0.274023</td>\n",
       "      <td>-9.003208</td>\n",
       "      <td>-17.048237</td>\n",
       "      <td>-21.399697</td>\n",
       "      <td>-16.263466</td>\n",
       "      <td>3.396780</td>\n",
       "      <td>-12.284809</td>\n",
       "      <td>0.908476</td>\n",
       "      <td>-10.026655</td>\n",
       "      <td>-4.254287</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>10.219182</td>\n",
       "      <td>1.742702</td>\n",
       "      <td>1.002258</td>\n",
       "      <td>-2.806073</td>\n",
       "      <td>-14.951893</td>\n",
       "      <td>-21.500796</td>\n",
       "      <td>-13.495337</td>\n",
       "      <td>-9.305933</td>\n",
       "      <td>-7.441661</td>\n",
       "      <td>-1.825364</td>\n",
       "      <td>-8.279968</td>\n",
       "      <td>0.422083</td>\n",
       "      <td>-3.959009</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>13.876193</td>\n",
       "      <td>1.004525</td>\n",
       "      <td>-10.143044</td>\n",
       "      <td>4.769317</td>\n",
       "      <td>-25.268881</td>\n",
       "      <td>-12.703184</td>\n",
       "      <td>-17.730170</td>\n",
       "      <td>-6.543746</td>\n",
       "      <td>-15.888943</td>\n",
       "      <td>-2.577060</td>\n",
       "      <td>-9.861565</td>\n",
       "      <td>1.551331</td>\n",
       "      <td>-3.312003</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>12.510054</td>\n",
       "      <td>-3.575093</td>\n",
       "      <td>-4.913943</td>\n",
       "      <td>-6.122454</td>\n",
       "      <td>-3.552648</td>\n",
       "      <td>-9.513949</td>\n",
       "      <td>-10.869105</td>\n",
       "      <td>-8.880711</td>\n",
       "      <td>1.368740</td>\n",
       "      <td>-5.230862</td>\n",
       "      <td>-0.560520</td>\n",
       "      <td>-4.260071</td>\n",
       "      <td>-8.647045</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>16.165702</td>\n",
       "      <td>7.492517</td>\n",
       "      <td>-18.944081</td>\n",
       "      <td>-11.769776</td>\n",
       "      <td>0.360446</td>\n",
       "      <td>-10.676715</td>\n",
       "      <td>-10.353008</td>\n",
       "      <td>7.848967</td>\n",
       "      <td>0.781196</td>\n",
       "      <td>-0.323822</td>\n",
       "      <td>-6.596552</td>\n",
       "      <td>-6.907072</td>\n",
       "      <td>-4.249575</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17375</th>\n",
       "      <td>17375</td>\n",
       "      <td>12.577980</td>\n",
       "      <td>0.449793</td>\n",
       "      <td>8.464243</td>\n",
       "      <td>13.715499</td>\n",
       "      <td>-1.113134</td>\n",
       "      <td>2.922420</td>\n",
       "      <td>-6.761397</td>\n",
       "      <td>2.240919</td>\n",
       "      <td>4.728756</td>\n",
       "      <td>-3.520493</td>\n",
       "      <td>6.343505</td>\n",
       "      <td>1.306161</td>\n",
       "      <td>-3.577430</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17376</th>\n",
       "      <td>17376</td>\n",
       "      <td>10.832928</td>\n",
       "      <td>5.712939</td>\n",
       "      <td>6.460498</td>\n",
       "      <td>0.479978</td>\n",
       "      <td>-9.827152</td>\n",
       "      <td>-8.491606</td>\n",
       "      <td>-16.057971</td>\n",
       "      <td>-7.821987</td>\n",
       "      <td>-7.278880</td>\n",
       "      <td>-8.225414</td>\n",
       "      <td>-1.474205</td>\n",
       "      <td>-10.781243</td>\n",
       "      <td>-8.982661</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17377</th>\n",
       "      <td>17377</td>\n",
       "      <td>15.653813</td>\n",
       "      <td>-6.198936</td>\n",
       "      <td>-0.756519</td>\n",
       "      <td>-0.182128</td>\n",
       "      <td>-5.058881</td>\n",
       "      <td>-12.113604</td>\n",
       "      <td>-15.959658</td>\n",
       "      <td>-1.984598</td>\n",
       "      <td>-3.341601</td>\n",
       "      <td>-5.365412</td>\n",
       "      <td>0.485510</td>\n",
       "      <td>-4.278507</td>\n",
       "      <td>-10.961661</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17378</th>\n",
       "      <td>17378</td>\n",
       "      <td>14.656198</td>\n",
       "      <td>2.442599</td>\n",
       "      <td>-1.116128</td>\n",
       "      <td>-3.430590</td>\n",
       "      <td>-9.892589</td>\n",
       "      <td>-16.597521</td>\n",
       "      <td>-17.447496</td>\n",
       "      <td>-1.346676</td>\n",
       "      <td>-11.960978</td>\n",
       "      <td>-9.395277</td>\n",
       "      <td>-5.763173</td>\n",
       "      <td>-7.353033</td>\n",
       "      <td>-8.212849</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17379</th>\n",
       "      <td>17379</td>\n",
       "      <td>10.457680</td>\n",
       "      <td>0.415493</td>\n",
       "      <td>-3.733206</td>\n",
       "      <td>5.588005</td>\n",
       "      <td>3.902790</td>\n",
       "      <td>-2.029628</td>\n",
       "      <td>-11.010215</td>\n",
       "      <td>-17.936656</td>\n",
       "      <td>-7.469568</td>\n",
       "      <td>-7.055387</td>\n",
       "      <td>-4.272895</td>\n",
       "      <td>-6.074557</td>\n",
       "      <td>-4.459090</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>17380 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Unnamed: 0          0         1          2          3          4  \\\n",
       "0               0  14.443077 -8.036094  -4.077253  -0.274023  -9.003208   \n",
       "1               1  10.219182  1.742702   1.002258  -2.806073 -14.951893   \n",
       "2               2  13.876193  1.004525 -10.143044   4.769317 -25.268881   \n",
       "3               3  12.510054 -3.575093  -4.913943  -6.122454  -3.552648   \n",
       "4               4  16.165702  7.492517 -18.944081 -11.769776   0.360446   \n",
       "...           ...        ...       ...        ...        ...        ...   \n",
       "17375       17375  12.577980  0.449793   8.464243  13.715499  -1.113134   \n",
       "17376       17376  10.832928  5.712939   6.460498   0.479978  -9.827152   \n",
       "17377       17377  15.653813 -6.198936  -0.756519  -0.182128  -5.058881   \n",
       "17378       17378  14.656198  2.442599  -1.116128  -3.430590  -9.892589   \n",
       "17379       17379  10.457680  0.415493  -3.733206   5.588005   3.902790   \n",
       "\n",
       "               5          6          7          8          9        10  \\\n",
       "0     -17.048237 -21.399697 -16.263466   3.396780 -12.284809  0.908476   \n",
       "1     -21.500796 -13.495337  -9.305933  -7.441661  -1.825364 -8.279968   \n",
       "2     -12.703184 -17.730170  -6.543746 -15.888943  -2.577060 -9.861565   \n",
       "3      -9.513949 -10.869105  -8.880711   1.368740  -5.230862 -0.560520   \n",
       "4     -10.676715 -10.353008   7.848967   0.781196  -0.323822 -6.596552   \n",
       "...          ...        ...        ...        ...        ...       ...   \n",
       "17375   2.922420  -6.761397   2.240919   4.728756  -3.520493  6.343505   \n",
       "17376  -8.491606 -16.057971  -7.821987  -7.278880  -8.225414 -1.474205   \n",
       "17377 -12.113604 -15.959658  -1.984598  -3.341601  -5.365412  0.485510   \n",
       "17378 -16.597521 -17.447496  -1.346676 -11.960978  -9.395277 -5.763173   \n",
       "17379  -2.029628 -11.010215 -17.936656  -7.469568  -7.055387 -4.272895   \n",
       "\n",
       "              11         12  LABEL  \n",
       "0     -10.026655  -4.254287      0  \n",
       "1       0.422083  -3.959009      0  \n",
       "2       1.551331  -3.312003      0  \n",
       "3      -4.260071  -8.647045      0  \n",
       "4      -6.907072  -4.249575      0  \n",
       "...          ...        ...    ...  \n",
       "17375   1.306161  -3.577430      1  \n",
       "17376 -10.781243  -8.982661      1  \n",
       "17377  -4.278507 -10.961661      1  \n",
       "17378  -7.353033  -8.212849      1  \n",
       "17379  -6.074557  -4.459090      1  \n",
       "\n",
       "[17380 rows x 15 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv(DATASET_PATH)\n",
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length of train dataset: 13904\n",
      "Length of test dataset: 3476\n"
     ]
    }
   ],
   "source": [
    "# randomly shuffle dataframe\n",
    "\n",
    "train = data.sample(frac=0.8, random_state=RANDOM_STATE)\n",
    "test = data.drop(train.index)\n",
    "\n",
    "print(\"Length of train dataset:\", len(train))\n",
    "print(\"Length of test dataset:\", len(test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train shape: (13904, 13)\n",
      "y_train shape: (13904,)\n"
     ]
    }
   ],
   "source": [
    "x_train = np.array(train)[:, 1:-1]\n",
    "y_train = np.array(train)[:, -1]\n",
    "\n",
    "print(\"x_train shape:\", x_train.shape)\n",
    "print(\"y_train shape:\", y_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_test shape: (3476, 13)\n",
      "y_test shape: (3476,)\n"
     ]
    }
   ],
   "source": [
    "x_test = np.array(test)[:, 1:-1]\n",
    "y_test = np.array(test)[:, -1]\n",
    "\n",
    "print(\"x_test shape:\", x_test.shape)\n",
    "print(\"y_test shape:\", y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "scaler = scaler.fit(x_train)\n",
    "\n",
    "x_train = scaler.transform(x_train)\n",
    "x_test = scaler.transform(x_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training And Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# hyper parameters\n",
    "\n",
    "LEARNING_RATE_INIT = 0.0001\n",
    "ALPHA = 0.012\n",
    "HIDDEN_LAYERS = (1024, 512, 512, 256, 32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MLPClassifier(alpha=0.012, hidden_layer_sizes=(1024, 512, 512, 256, 32),\n",
       "              learning_rate_init=0.0001)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = MLPClassifier(hidden_layer_sizes=(HIDDEN_LAYERS), alpha=ALPHA, learning_rate_init=LEARNING_RATE_INIT)\n",
    "model.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss = model.loss_curve_\n",
    "\n",
    "plt.style.use(\"ggplot\")\n",
    "plt.figure(figsize=(10, 10))\n",
    "plt.plot(range(len(loss)), loss)\n",
    "plt.xlabel(\"Iterations\")\n",
    "plt.ylabel(\"Cost\")\n",
    "plt.title(\"Cost against iterations for best Estimator\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5 Fold CV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "nn = MLPClassifier(hidden_layer_sizes=(HIDDEN_LAYERS), alpha=ALPHA, learning_rate_init=LEARNING_RATE_INIT)\n",
    "nn.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\neural_network\\_multilayer_perceptron.py:619: UserWarning: Training interrupted by user.\n",
      "  warnings.warn(\"Training interrupted by user.\")\n"
     ]
    }
   ],
   "source": [
    "cv_scores = cross_val_score(nn, x_train, y_train, cv=CROSS_VALIDATION_SETS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = [x for x in range(1, CROSS_VALIDATION_SETS + 1)]\n",
    "\n",
    "plot(x, cv_scores, 'CV Sets', 'Accuracy', '5-Fold CV Accuracy Plot')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing and Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "     English       0.77      0.78      0.77      1060\n",
      "        Urdu       0.76      0.77      0.76      1020\n",
      "       Mixed       0.84      0.82      0.83      1396\n",
      "\n",
      "    accuracy                           0.79      3476\n",
      "   macro avg       0.79      0.79      0.79      3476\n",
      "weighted avg       0.79      0.79      0.79      3476\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, preds, target_names=LABEL_NAMES))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "conf = confusion_matrix(y_test, preds)\n",
    "matrix_plotter = ConfusionMatrixDisplay(conf, display_labels=LABEL_NAMES)\n",
    "\n",
    "\n",
    "a = matrix_plotter.plot(ax=None)\n",
    "a.ax_.grid()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "b3ba2566441a7c06988d0923437866b63cedc61552a5af99d1f4fb67d367b25f"
  },
  "kernelspec": {
   "display_name": "Python 3.8.8 64-bit ('base': conda)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}