NaiveBayes.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Naive Bayes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Importing necessary libraries**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Loading the preprocessed dataset**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "reviews_df = pd.read_csv(\"preprocessed_data_max.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "label = \"is_bad_review\"\n",
    "ignore_cols = [label, \"review\", \"review_clean\", \"compound\"]\n",
    "features = [c for c in reviews_df.columns if c not in ignore_cols]\n",
    "\n",
    "\n",
    "X = reviews_df[features]\n",
    "y = reviews_df[label]\n",
    " \n",
    "n_features = X.shape[1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Making a split for Validation Dataset**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(X, y, random_state = 42, test_size = 0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Creating custom MNB Class**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class MNB(object):\n",
    "    \n",
    "    def _init_(self):\n",
    "        self.probailities = None\n",
    "        self.parameters = None\n",
    "        self.class_labels = None\n",
    "    \n",
    "    def fit(self, X, y, alpha=1.0):\n",
    "                \n",
    "        self.class_labels = np.unique(y)\n",
    "        self.parameters = np.zeros(shape = (X.shape[1], len(self.class_labels)))\n",
    "        self.probabilities = np.zeros(shape = (len(self.class_labels),))\n",
    "        \n",
    "        for ix,label in enumerate(self.class_labels):\n",
    "            selected_class = (y == label)\n",
    "    \n",
    "            token_counts_in_label = []\n",
    "            for i in range(X.shape[1]):\n",
    "                temp = 0\n",
    "                temp += np.sum(X[selected_class, i]) + alpha     \n",
    "                token_counts_in_label.append(temp)\n",
    "            \n",
    "            total_words_in_label = X.shape[1] * alpha\n",
    "            total_words_in_label += np.sum(X[selected_class,:])\n",
    "            self.parameters[:, ix] = token_counts_in_label / total_words_in_label\n",
    "            self.probabilities[ix] = np.sum(selected_class)/len(y)\n",
    "    \n",
    "    def predict_log_likelihood(self, X):\n",
    "\n",
    "        log_parameters = np.log(self.parameters)\n",
    "        log_likelihoods = np.dot(X, log_parameters)\n",
    "        for i in range(len(log_likelihoods)):\n",
    "            for j in range(len(self.probabilities)):\n",
    "                log_likelihoods[i][j] += np.log(self.probabilities[j])\n",
    "        return log_likelihoods\n",
    "            \n",
    "    def predict(self, X):\n",
    "        print(self.probabilities)\n",
    "        log_likelihoods = self.predict_log_likelihood(X)\n",
    "        label = np.argmax(log_likelihoods, axis=1)\n",
    "        pred_y = [self.class_labels[index] for index in label]\n",
    "        \n",
    "        return pred_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "nbcustom = MNB()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "nbcustom.fit(X_train.to_numpy(), y_train.to_numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.59630682 0.40369318]\n"
     ]
    }
   ],
   "source": [
    "y1 = nbcustom.predict(X_train.to_numpy())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Training Results**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy : 0.8266477272727273\n",
      "AUC Score : 0.8134901543209979\n",
      "F1 Score : 0.7763196480938417\n",
      "Precision : 0.8101759755164499\n",
      "Recall : 0.7451794510907811\n"
     ]
    }
   ],
   "source": [
    "print(f'Accuracy : {accuracy_score(y_train, y1)}')\n",
    "print(f'AUC Score : {roc_auc_score(y_train, y1)}')\n",
    "print(f'F1 Score : {f1_score(y_train, y1)}')\n",
    "print(f'Precision : {precision_score(y_train, y1)}')\n",
    "print(f'Recall : {recall_score(y_train, y1)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.59630682 0.40369318]\n"
     ]
    }
   ],
   "source": [
    "y2 = nbcustom.predict(X_val.to_numpy())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Validation Results**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy : 0.8232954545454545\n",
      "AUC Score : 0.811782617531265\n",
      "F1 Score : 0.7766767198046819\n",
      "Precision : 0.8090963494913226\n",
      "Recall : 0.7467550400441867\n"
     ]
    }
   ],
   "source": [
    "print(f'Accuracy : {accuracy_score(y_val, y2)}')\n",
    "print(f'AUC Score : {roc_auc_score(y_val, y2)}')\n",
    "print(f'F1 Score : {f1_score(y_val, y2)}')\n",
    "print(f'Precision : {precision_score(y_val, y2)}')\n",
    "print(f'Recall : {recall_score(y_val, y2)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Training the dataset on Multinomial Naive Bayes Algorithm from sklearn library**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.naive_bayes import MultinomialNB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultinomialNB()"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nb = MultinomialNB()\n",
    "nb.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y1 = nb.predict(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score, roc_auc_score, roc_curve, f1_score,  precision_score, recall_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Training Scores**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy : 0.8266477272727273\n",
      "AUC Score : 0.8134901543209979\n",
      "F1 Score : 0.7763196480938417\n",
      "Precision : 0.8101759755164499\n",
      "Recall : 0.7451794510907811\n"
     ]
    }
   ],
   "source": [
    "print(f'Accuracy : {accuracy_score(y_train, y1)}')\n",
    "print(f'AUC Score : {roc_auc_score(y_train, y1)}')\n",
    "print(f'F1 Score : {f1_score(y_train, y1)}')\n",
    "print(f'Precision : {precision_score(y_train, y1)}')\n",
    "print(f'Recall : {recall_score(y_train, y1)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "y2 = nb.predict(X_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Validation scores**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy : 0.8232954545454545\n",
      "AUC Score : 0.811782617531265\n",
      "F1 Score : 0.7766767198046819\n",
      "Precision : 0.8090963494913226\n",
      "Recall : 0.7467550400441867\n"
     ]
    }
   ],
   "source": [
    "print(f'Accuracy : {accuracy_score(y_val, y2)}')\n",
    "print(f'AUC Score : {roc_auc_score(y_val, y2)}')\n",
    "print(f'F1 Score : {f1_score(y_val, y2)}')\n",
    "print(f'Precision : {precision_score(y_val, y2)}')\n",
    "print(f'Recall : {recall_score(y_val, y2)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Comparison of Training anf Validation Results**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,5))\n",
    "barWidth = 0.3\n",
    "\n",
    "bars1 = [accuracy_score(y_train, y1), roc_auc_score(y_train, y1), f1_score(y_train, y1), precision_score(y_train, y1), recall_score(y_train, y1)]\n",
    "bars2 = [accuracy_score(y_val, y2), roc_auc_score(y_val, y2), f1_score(y_val, y2), precision_score(y_val, y2), recall_score(y_val, y2)]\n",
    " \n",
    "r1 = np.arange(len(bars1))\n",
    "r2 = [x + barWidth for x in r1]\n",
    " \n",
    "plt.bar(r1, bars1, width = barWidth, color = (0.3,0.5,0.4,0.6), capsize=7, label='Training')\n",
    " \n",
    "plt.bar(r2, bars2, width = barWidth, color = (0.3,0.9,0.4,0.6), capsize=7, label='Validation')\n",
    "\n",
    "plt.xticks([r + barWidth for r in range(len(bars1))], ['Accuracy', 'ROC-AUC', 'F1 Score', 'Precision', 'Recall'])\n",
    "plt.ylabel('Score')\n",
    "plt.legend(loc = 'best')\n",
    " \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Saving stored model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['nb.pkl']"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import joblib\n",
    "joblib.dump(nb, 'nb.pkl', compress=1) "
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}