Naive Bayes Classifier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "e7fced6e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          package_name                                             review  \\\n",
      "0  com.facebook.katana   privacy at least put some option appear offli...   \n",
      "1  com.facebook.katana   messenger issues ever since the last update, ...   \n",
      "2  com.facebook.katana   profile any time my wife or anybody has more ...   \n",
      "3  com.facebook.katana   the new features suck for those of us who don...   \n",
      "4  com.facebook.katana   forced reload on uploading pic on replying co...   \n",
      "\n",
      "   polarity  \n",
      "0         0  \n",
      "1         0  \n",
      "2         0  \n",
      "3         0  \n",
      "4         0  \n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "data = pd.read_csv('Datasets/google_play_store_apps_reviews_training.csv')\n",
    "print(data.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "cd20f248",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Pre-process Data\n",
    "def preprocess_data(data):\n",
    "    # Remove package name as it's not relevant\n",
    "    data = data.drop('package_name', axis=1)\n",
    "    # Convert text to lowercase\n",
    "    data['review'] = data['review'].str.strip().str.lower()\n",
    "    return data\n",
    "data = preprocess_data(data)\n",
    "\n",
    "\n",
    "x = data['review']\n",
    "y = data['polarity']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ae506f2e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Vectorize text reviews to numbers\n",
    "vec = CountVectorizer(stop_words='english')\n",
    "x = vec.fit_transform(x).toarray()\n",
    "# Split into training and testing data\n",
    "x_train, x_test, y_train, y_test = train_test_split(x,y, test_size=0.25, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c8e8177a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MultinomialNB()\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.9327354260089686"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Model Generation\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "model = MultinomialNB()\n",
    "print(model.fit(x, y))\n",
    "y_pred =model.predict(x_test)\n",
    "model.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9a9789e0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[152   6]\n",
      " [  9  56]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      0.96      0.95       158\n",
      "           1       0.90      0.86      0.88        65\n",
      "\n",
      "    accuracy                           0.93       223\n",
      "   macro avg       0.92      0.91      0.92       223\n",
      "weighted avg       0.93      0.93      0.93       223\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([1], dtype=int64)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))\n",
    "model.predict(vec.transform(['Love this app simply awesome!']))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd7a4871",
   "metadata": {},
   "source": [
    "The confusion matrix and classification report generated by the code indicate that the model has a high overall accuracy of 93%, with an accuracy of 94% for negative reviews (polarity 0) and 90% for positive reviews (polarity 1).\n",
    "\n",
    "The confusion matrix shows that out of the 158 negative reviews in the test set, the model correctly predicted 152 as negative and incorrectly predicted 6 as positive. Out of the 65 positive reviews in the test set, the model correctly predicted 56 as positive and incorrectly predicted 9 as negative."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a2b0120",
   "metadata": {},
   "source": [
    "The precision for negative reviews is 94% and for positive reviews is 90%, which means that when the model predicts a review as negative or positive, it is correct 94% and 90% of the time, respectively. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5472988e",
   "metadata": {},
   "source": [
    "The recall for negative reviews is 96% and for positive reviews is 86%, which means that the model correctly identifies 96% and 86% of the negative and positive reviews, respectively. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a9abdf9",
   "metadata": {},
   "source": [
    "The F1-score for negative reviews is 95% and for positive reviews is 88%, which is a weighted harmonic mean of precision and recall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "31962275",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Predict for Custom Input:\n",
    "def expression_check(prediction_input):\n",
    "    if prediction_input == 0:\n",
    "        print(\"Input statement has Negative Sentiment.\")\n",
    "    elif prediction_input == 1:\n",
    "        print(\"Input statement has Positive Sentiment.\")\n",
    "    else:\n",
    "        print(\"Invalid Statement.\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ca11ed11",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input statement has Negative Sentiment.\n",
      "Input statement has Positive Sentiment.\n"
     ]
    }
   ],
   "source": [
    "# function to take the input statement and perform the same transformations we did earlier\n",
    "def sentiment_predictor(input):\n",
    "    transformed_input = vec.transform(input)\n",
    "    prediction = model.predict(transformed_input)\n",
    "    expression_check(prediction)\n",
    "\n",
    "input1 = [\"Worst support. I have given 2 stars because changes to be done.\"]\n",
    "input2 = [\"Its working not nicely.\"]\n",
    "\n",
    "\n",
    "sentiment_predictor(input1)\n",
    "sentiment_predictor(input2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8b091743",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}