Sentiment_Classificationa.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c56bb893",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing Libraries\n",
    "import pandas as pd\n",
    "import preprocess\n",
    "import numpy as np"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4fcc25e4",
   "metadata": {},
   "source": [
    "Conda deactivate\n",
    "Streamlit run streamlit_app.py \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "dfe00875-6a5c-4807-810b-88f5699f2956",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\Dell\\AppData\\Local\\Programs\\Python\\Python310\\lib\\site-packages\\pandas\\core\\strings\\object_array.py:158: FutureWarning: Possible nested set at position 1\n",
      "  pat = re.compile(pat, flags=flags)\n"
     ]
    },
    {
     "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>Text</th>\n",
       "      <th>Sentiment</th>\n",
       "      <th>Text_Cleaned</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Kickers on my watchlist XIDE TIT SOQ PNK CPW B...</td>\n",
       "      <td>1</td>\n",
       "      <td>kickers on my watchlist xide tit soq pnk cpw b...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>user: AAP MOVIE. 55% return for the FEA/GEED i...</td>\n",
       "      <td>1</td>\n",
       "      <td>aap movie 55 percent return for the fea and ge...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>user I'd be afraid to short AMZN - they are lo...</td>\n",
       "      <td>1</td>\n",
       "      <td>id be afraid to short amzn to they are looking...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MNTA Over 12.00</td>\n",
       "      <td>1</td>\n",
       "      <td>mnta over 12.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>OI  Over 21.37</td>\n",
       "      <td>1</td>\n",
       "      <td>oi over 21.37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5786</th>\n",
       "      <td>Industry body CII said #discoms are likely to ...</td>\n",
       "      <td>-1</td>\n",
       "      <td>industry body cii said are likely to suffer a ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5787</th>\n",
       "      <td>#Gold prices slip below Rs 46,000 as #investor...</td>\n",
       "      <td>-1</td>\n",
       "      <td>prices slip below rs 46000 as book profits ami...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5788</th>\n",
       "      <td>Workers at Bajaj Auto have agreed to a 10% wag...</td>\n",
       "      <td>1</td>\n",
       "      <td>workers at bajaj auto have agreed to a 10 perc...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5789</th>\n",
       "      <td>#Sharemarket LIVE: Sensex off day’s high, up 6...</td>\n",
       "      <td>1</td>\n",
       "      <td>live sensex off day’s high up 600 points tests...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5790</th>\n",
       "      <td>#Sensex, #Nifty climb off day's highs, still u...</td>\n",
       "      <td>1</td>\n",
       "      <td>climb off days highs still up 2 percent key fa...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5791 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   Text  Sentiment  \\\n",
       "0     Kickers on my watchlist XIDE TIT SOQ PNK CPW B...          1   \n",
       "1     user: AAP MOVIE. 55% return for the FEA/GEED i...          1   \n",
       "2     user I'd be afraid to short AMZN - they are lo...          1   \n",
       "3                                     MNTA Over 12.00            1   \n",
       "4                                      OI  Over 21.37            1   \n",
       "...                                                 ...        ...   \n",
       "5786  Industry body CII said #discoms are likely to ...         -1   \n",
       "5787  #Gold prices slip below Rs 46,000 as #investor...         -1   \n",
       "5788  Workers at Bajaj Auto have agreed to a 10% wag...          1   \n",
       "5789  #Sharemarket LIVE: Sensex off day’s high, up 6...          1   \n",
       "5790  #Sensex, #Nifty climb off day's highs, still u...          1   \n",
       "\n",
       "                                           Text_Cleaned  \n",
       "0     kickers on my watchlist xide tit soq pnk cpw b...  \n",
       "1     aap movie 55 percent return for the fea and ge...  \n",
       "2     id be afraid to short amzn to they are looking...  \n",
       "3                                       mnta over 12.00  \n",
       "4                                         oi over 21.37  \n",
       "...                                                 ...  \n",
       "5786  industry body cii said are likely to suffer a ...  \n",
       "5787  prices slip below rs 46000 as book profits ami...  \n",
       "5788  workers at bajaj auto have agreed to a 10 perc...  \n",
       "5789  live sensex off day’s high up 600 points tests...  \n",
       "5790  climb off days highs still up 2 percent key fa...  \n",
       "\n",
       "[5791 rows x 3 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "# Read in labeled stock sentiment data\n",
    "data = pd.read_csv('train/stock_data.csv')\n",
    "\n",
    "# Preprocess tweets for NLP analysis\n",
    "data = preprocess.Preprocess_Tweets(data)\n",
    "display(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8d5d86e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "#import nltk\n",
    "#nltk.download('vader_lexicon')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "4ffe33ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk\n",
    "#nltk.download('vader_lexicon')\n",
    "from nltk.sentiment.vader import SentimentIntensityAnalyzer\n",
    "\n",
    "# Prepare Vader sentiment analyzer\n",
    "sid = SentimentIntensityAnalyzer()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "fe113335",
   "metadata": {},
   "source": [
    "Vedar Model Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "82c4cd4e-d6f6-4907-b5e4-fac8e79d2c2c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Vader Accuracy: 66.48 % \n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Predict sentiment with Vader classifier\n",
    "data['Vader_Scores'] = data['Text_Cleaned'].apply(lambda score: sid.polarity_scores(score)['compound'])\n",
    "data['Vader_Prediction'] = data['Vader_Scores'].apply(lambda score: 1 if score >=0 else -1)\n",
    "\n",
    "# Print Vader sentiment accuracy\n",
    "print('Vader Accuracy:', round((len(data[data['Sentiment']==data['Vader_Prediction']])/len(data)) *100, 2), '%', '\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "f7809401",
   "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>Text</th>\n",
       "      <th>Sentiment</th>\n",
       "      <th>Text_Cleaned</th>\n",
       "      <th>Vader_Scores</th>\n",
       "      <th>Vader_Prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Kickers on my watchlist XIDE TIT SOQ PNK CPW B...</td>\n",
       "      <td>1</td>\n",
       "      <td>kickers on my watchlist xide tit soq pnk cpw b...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>user: AAP MOVIE. 55% return for the FEA/GEED i...</td>\n",
       "      <td>1</td>\n",
       "      <td>aap movie 55 percent return for the fea and ge...</td>\n",
       "      <td>0.6249</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>user I'd be afraid to short AMZN - they are lo...</td>\n",
       "      <td>1</td>\n",
       "      <td>id be afraid to short amzn to they are looking...</td>\n",
       "      <td>0.3612</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MNTA Over 12.00</td>\n",
       "      <td>1</td>\n",
       "      <td>mnta over 12.00</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>OI  Over 21.37</td>\n",
       "      <td>1</td>\n",
       "      <td>oi over 21.37</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5786</th>\n",
       "      <td>Industry body CII said #discoms are likely to ...</td>\n",
       "      <td>-1</td>\n",
       "      <td>industry body cii said are likely to suffer a ...</td>\n",
       "      <td>-0.7003</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5787</th>\n",
       "      <td>#Gold prices slip below Rs 46,000 as #investor...</td>\n",
       "      <td>-1</td>\n",
       "      <td>prices slip below rs 46000 as book profits ami...</td>\n",
       "      <td>0.0258</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5788</th>\n",
       "      <td>Workers at Bajaj Auto have agreed to a 10% wag...</td>\n",
       "      <td>1</td>\n",
       "      <td>workers at bajaj auto have agreed to a 10 perc...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5789</th>\n",
       "      <td>#Sharemarket LIVE: Sensex off day’s high, up 6...</td>\n",
       "      <td>1</td>\n",
       "      <td>live sensex off day’s high up 600 points tests...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5790</th>\n",
       "      <td>#Sensex, #Nifty climb off day's highs, still u...</td>\n",
       "      <td>1</td>\n",
       "      <td>climb off days highs still up 2 percent key fa...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5791 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   Text  Sentiment  \\\n",
       "0     Kickers on my watchlist XIDE TIT SOQ PNK CPW B...          1   \n",
       "1     user: AAP MOVIE. 55% return for the FEA/GEED i...          1   \n",
       "2     user I'd be afraid to short AMZN - they are lo...          1   \n",
       "3                                     MNTA Over 12.00            1   \n",
       "4                                      OI  Over 21.37            1   \n",
       "...                                                 ...        ...   \n",
       "5786  Industry body CII said #discoms are likely to ...         -1   \n",
       "5787  #Gold prices slip below Rs 46,000 as #investor...         -1   \n",
       "5788  Workers at Bajaj Auto have agreed to a 10% wag...          1   \n",
       "5789  #Sharemarket LIVE: Sensex off day’s high, up 6...          1   \n",
       "5790  #Sensex, #Nifty climb off day's highs, still u...          1   \n",
       "\n",
       "                                           Text_Cleaned  Vader_Scores  \\\n",
       "0     kickers on my watchlist xide tit soq pnk cpw b...        0.0000   \n",
       "1     aap movie 55 percent return for the fea and ge...        0.6249   \n",
       "2     id be afraid to short amzn to they are looking...        0.3612   \n",
       "3                                       mnta over 12.00        0.0000   \n",
       "4                                         oi over 21.37        0.0000   \n",
       "...                                                 ...           ...   \n",
       "5786  industry body cii said are likely to suffer a ...       -0.7003   \n",
       "5787  prices slip below rs 46000 as book profits ami...        0.0258   \n",
       "5788  workers at bajaj auto have agreed to a 10 perc...        0.0000   \n",
       "5789  live sensex off day’s high up 600 points tests...        0.0000   \n",
       "5790  climb off days highs still up 2 percent key fa...        0.0000   \n",
       "\n",
       "      Vader_Prediction  \n",
       "0                    1  \n",
       "1                    1  \n",
       "2                    1  \n",
       "3                    1  \n",
       "4                    1  \n",
       "...                ...  \n",
       "5786                -1  \n",
       "5787                 1  \n",
       "5788                 1  \n",
       "5789                 1  \n",
       "5790                 1  \n",
       "\n",
       "[5791 rows x 5 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(data)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4fddd11b",
   "metadata": {},
   "source": [
    "Visualizing the Accuracy of Vedar Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "9525ceae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "# Plot Vader sentiment accuracy\n",
    "import matplotlib.pyplot as plt\n",
    "labels = ['Correct Predictions', 'Incorrect Predictions']\n",
    "sizes = [len(data[data['Sentiment'] == data['Vader_Prediction']]), len(data[data['Sentiment'] != data['Vader_Prediction']])]\n",
    "colors = ['#66BB6A', '#EF5350']\n",
    "\n",
    "plt.pie(sizes, labels=labels, colors=colors, autopct='%1.1f%%', startangle=90)\n",
    "plt.axis('equal')\n",
    "\n",
    "plt.title('Vader Sentiment Accuracy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "fa772f3d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\Dell\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import nltk\n",
    "nltk.download(\"stopwords\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "15a40321",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', \"you're\", \"you've\", \"you'll\", \"you'd\", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', \"she's\", 'her', 'hers', 'herself', 'it', \"it's\", 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', \"that'll\", 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', \"don't\", 'should', \"should've\", 'now', 'd', 'll', 'm', 'o', 're', 've', 'y', 'ain', 'aren', \"aren't\", 'couldn', \"couldn't\", 'didn', \"didn't\", 'doesn', \"doesn't\", 'hadn', \"hadn't\", 'hasn', \"hasn't\", 'haven', \"haven't\", 'isn', \"isn't\", 'ma', 'mightn', \"mightn't\", 'mustn', \"mustn't\", 'needn', \"needn't\", 'shan', \"shan't\", 'shouldn', \"shouldn't\", 'wasn', \"wasn't\", 'weren', \"weren't\", 'won', \"won't\", 'wouldn', \"wouldn't\"]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\Dell\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import nltk\n",
    "from nltk.corpus import stopwords\n",
    "\n",
    "# Download the stopwords if not already downloaded\n",
    "nltk.download('stopwords')\n",
    "\n",
    "# Print the list of stopwords for the English language\n",
    "print(stopwords.words('english'))\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "c6eea157",
   "metadata": {},
   "source": [
    "Stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "740d48ce-4491-4160-8bad-b3e2a0ab9ebd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'them', 'these', 'such', 'just', 'youre', 'hadnt', 'wouldnt', 'wouldn', 'out', 'if', 'thatll', 'mightn', 'isn', 'should', 'in', 'who', 'theirs', 'couldnt', 'more', 't', 'me', 'youd', 'before', 'has', 'his', 'did', 'where', 'hers', 'it', 'were', 'for', 'don', 'be', 'on', 'as', 'while', 'most', 've', 'doesnt', 'ourselves', 'and', 'he', 'off', 'with', 'some', 'nor', 'here', 'doing', 'of', 're', 'didnt', 'havent', 'an', 'your', 'a', 'was', 'too', 'they', 'we', 'werent', 'now', 'wont', 'her', 'when', 'having', 'myself', 'both', 'which', 'shant', 'all', 'mustnt', 'again', 'had', 'are', 'shouldve', 'have', 'hadn', 'any', 'between', 'yours', 'themselves', 'him', 'i', 'mightnt', 'y', 'being', 'own', 'haven', 'neednt', 'himself', 'do', 'mustn', 'than', 'through', 'she', 'arent', 'its', 'itself', 'herself', 'youll', 'about', 'shes', 'shouldn', 'wasnt', 'whom', 'aren', 'very', 'against', 'once', 'ma', 'can', 'that', 'there', 'dont', 'but', 'so', 'o', 'is', 'our', 'how', 'been', 'because', 'into', 'only', 'couldn', 'hasn', 'ours', 'shouldnt', 'few', 'or', 'why', 'weren', 'doesn', 'no', 'wasn', 'm', 'what', 'hasnt', 'further', 'this', 'their', 'the', 'then', 'after', 'until', 'needn', 'yourselves', 'by', 'you', 'to', 'from', 'ain', 'yourself', 'am', 'will', 'won', 's', 'those', 'my', 'does', 'during', 'd', 'shan', 'other', 'at', 'same', 'didn', 'isnt', 'youve', 'each', 'll'}\n"
     ]
    },
    {
     "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>Text</th>\n",
       "      <th>Sentiment</th>\n",
       "      <th>Text_Cleaned</th>\n",
       "      <th>Vader_Scores</th>\n",
       "      <th>Vader_Prediction</th>\n",
       "      <th>Text_Processed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Kickers on my watchlist XIDE TIT SOQ PNK CPW B...</td>\n",
       "      <td>1</td>\n",
       "      <td>kickers on my watchlist xide tit soq pnk cpw b...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>kickers watchlist xide tit soq pnk cpw bpz aj ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>user: AAP MOVIE. 55% return for the FEA/GEED i...</td>\n",
       "      <td>1</td>\n",
       "      <td>aap movie 55 percent return for the fea and ge...</td>\n",
       "      <td>0.6249</td>\n",
       "      <td>1</td>\n",
       "      <td>aap movie 55 percent return fea geed indicator...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>user I'd be afraid to short AMZN - they are lo...</td>\n",
       "      <td>1</td>\n",
       "      <td>id be afraid to short amzn to they are looking...</td>\n",
       "      <td>0.3612</td>\n",
       "      <td>1</td>\n",
       "      <td>id afraid short amzn looking like near monopol...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MNTA Over 12.00</td>\n",
       "      <td>1</td>\n",
       "      <td>mnta over 12.00</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>mnta over 12.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>OI  Over 21.37</td>\n",
       "      <td>1</td>\n",
       "      <td>oi over 21.37</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>oi over 21.37</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5786</th>\n",
       "      <td>Industry body CII said #discoms are likely to ...</td>\n",
       "      <td>-1</td>\n",
       "      <td>industry body cii said are likely to suffer a ...</td>\n",
       "      <td>-0.7003</td>\n",
       "      <td>-1</td>\n",
       "      <td>industry body cii said likely suffer net reven...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5787</th>\n",
       "      <td>#Gold prices slip below Rs 46,000 as #investor...</td>\n",
       "      <td>-1</td>\n",
       "      <td>prices slip below rs 46000 as book profits ami...</td>\n",
       "      <td>0.0258</td>\n",
       "      <td>1</td>\n",
       "      <td>prices slip below rs 46000 book profits amid l...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5788</th>\n",
       "      <td>Workers at Bajaj Auto have agreed to a 10% wag...</td>\n",
       "      <td>1</td>\n",
       "      <td>workers at bajaj auto have agreed to a 10 perc...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>workers bajaj auto agreed 10 percent wage cut ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5789</th>\n",
       "      <td>#Sharemarket LIVE: Sensex off day’s high, up 6...</td>\n",
       "      <td>1</td>\n",
       "      <td>live sensex off day’s high up 600 points tests...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>live sensex day’s high up 600 points tests 920...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5790</th>\n",
       "      <td>#Sensex, #Nifty climb off day's highs, still u...</td>\n",
       "      <td>1</td>\n",
       "      <td>climb off days highs still up 2 percent key fa...</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>climb days highs still up 2 percent key factor...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5791 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   Text  Sentiment  \\\n",
       "0     Kickers on my watchlist XIDE TIT SOQ PNK CPW B...          1   \n",
       "1     user: AAP MOVIE. 55% return for the FEA/GEED i...          1   \n",
       "2     user I'd be afraid to short AMZN - they are lo...          1   \n",
       "3                                     MNTA Over 12.00            1   \n",
       "4                                      OI  Over 21.37            1   \n",
       "...                                                 ...        ...   \n",
       "5786  Industry body CII said #discoms are likely to ...         -1   \n",
       "5787  #Gold prices slip below Rs 46,000 as #investor...         -1   \n",
       "5788  Workers at Bajaj Auto have agreed to a 10% wag...          1   \n",
       "5789  #Sharemarket LIVE: Sensex off day’s high, up 6...          1   \n",
       "5790  #Sensex, #Nifty climb off day's highs, still u...          1   \n",
       "\n",
       "                                           Text_Cleaned  Vader_Scores  \\\n",
       "0     kickers on my watchlist xide tit soq pnk cpw b...        0.0000   \n",
       "1     aap movie 55 percent return for the fea and ge...        0.6249   \n",
       "2     id be afraid to short amzn to they are looking...        0.3612   \n",
       "3                                       mnta over 12.00        0.0000   \n",
       "4                                         oi over 21.37        0.0000   \n",
       "...                                                 ...           ...   \n",
       "5786  industry body cii said are likely to suffer a ...       -0.7003   \n",
       "5787  prices slip below rs 46000 as book profits ami...        0.0258   \n",
       "5788  workers at bajaj auto have agreed to a 10 perc...        0.0000   \n",
       "5789  live sensex off day’s high up 600 points tests...        0.0000   \n",
       "5790  climb off days highs still up 2 percent key fa...        0.0000   \n",
       "\n",
       "      Vader_Prediction                                     Text_Processed  \n",
       "0                    1  kickers watchlist xide tit soq pnk cpw bpz aj ...  \n",
       "1                    1  aap movie 55 percent return fea geed indicator...  \n",
       "2                    1  id afraid short amzn looking like near monopol...  \n",
       "3                    1                                    mnta over 12.00  \n",
       "4                    1                                      oi over 21.37  \n",
       "...                ...                                                ...  \n",
       "5786                -1  industry body cii said likely suffer net reven...  \n",
       "5787                 1  prices slip below rs 46000 book profits amid l...  \n",
       "5788                 1  workers bajaj auto agreed 10 percent wage cut ...  \n",
       "5789                 1  live sensex day’s high up 600 points tests 920...  \n",
       "5790                 1  climb days highs still up 2 percent key factor...  \n",
       "\n",
       "[5791 rows x 6 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import nltk\n",
    "#nltk.download(\"stopwords\")\n",
    "from nltk.corpus import stopwords\n",
    "\n",
    "# Create list of custom stop words to remove\n",
    "StopWords = set([s.replace(\"'\", '') for s in stopwords.words('english') if s not in ['not', 'up', 'down', 'above', 'below', 'under', 'over']])\n",
    "print(StopWords)\n",
    "\n",
    "# Remove stop words for tweets\n",
    "data['Text_Processed'] = data['Text_Cleaned'].apply(lambda s: \" \".join([word for word in s.split() if word not in StopWords]))\n",
    "data['Text_Processed'] = data['Text_Processed'].str.strip()\n",
    "\n",
    "display(data)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "d06c29c4",
   "metadata": {},
   "source": [
    "Importing Libraries for Naive Bayes Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "78d57a13",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.model_selection import StratifiedKFold, cross_val_score #for model accuracy\n",
    "from sklearn.naive_bayes import MultinomialNB #Model\n",
    "import matplotlib.pyplot as plt\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "14e9ddf5",
   "metadata": {},
   "source": [
    "Splitting Data into Training and Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ee63045b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split data into 80/20 train-test split\n",
    "train_pct = .8\n",
    "np.random.seed(1)  #ets the seed value for the random number generator\n",
    "idx = np.random.permutation(len(data)) #randomize data"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "68fa344f",
   "metadata": {},
   "source": [
    "Model Fitting and Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d987e58b",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = data['Text_Processed'].values[idx[:int(train_pct*len(data))]]\n",
    "y_train = data['Sentiment'].values[idx[:int(train_pct*len(data))]]\n",
    "\n",
    "y_train[y_train==-1] = 0 #Naive Bayes output values as 0 and 1, and write now we have 1 and -1\n",
    "\n",
    "X_test = data['Text_Processed'].values[idx[int(train_pct*len(data)):]]\n",
    "y_test = data['Sentiment'].values[idx[int(train_pct*len(data)):]]\n",
    "y_test[y_test==-1] = 0  #Naive Bayes output values as 0 and 1, and write now we have 1 and -1\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "a5030f7c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4632,) (4632,)\n",
      "(1159,) (1159,)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape, y_train.shape )\n",
    "print(X_test.shape, y_test.shape )"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "9d61e7df",
   "metadata": {},
   "source": [
    "Calculate TF-IDF for Naive Bayes classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "a61ddfc7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Initilize Vectorizer\n",
    "\n",
    "tf_idf = TfidfVectorizer(ngram_range=(1, 3),\n",
    "                         binary=True,\n",
    "                         smooth_idf=False)\n",
    "\n",
    "# Get TF-IDF for Train and Test data\n",
    "X_train_tfidf = tf_idf.fit_transform(X_train)\n",
    "X_test_tfidf = tf_idf.transform(X_test)\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "47c81d93",
   "metadata": {},
   "source": [
    "Model Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "ef8b54b7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best alpha:  1.0   |  Best Score:  83.04\n"
     ]
    }
   ],
   "source": [
    "# Define function to determine accuracy of model\n",
    "def get_auc_CV(model):\n",
    "    # Set KFold to shuffle data before the split\n",
    "    kf = StratifiedKFold(5, shuffle=True, random_state=1)\n",
    "\n",
    "    # Get AUC scores\n",
    "    auc = cross_val_score(model, X_train_tfidf, y_train, scoring=\"roc_auc\", cv=kf)\n",
    "\n",
    "    return auc.mean()\n",
    "\n",
    "# Initialize models with multiple alpha values to find best model\n",
    "#Alpha Values for different mdoels\n",
    "alphas = np.arange(1,10,0.1)\n",
    "models = [MultinomialNB(alpha=i) for i in alphas]\n",
    "\n",
    "# Find best performing model\n",
    "accs = []\n",
    "for model in models:\n",
    "    accs.append(get_auc_CV(model))\n",
    "    \n",
    "accs = np.array(accs)\n",
    "\n",
    "# Get best performing alpha value to continue with\n",
    "best_alpha = round(alphas[accs.argmax()], 1)\n",
    "\n",
    "# Print best alpha value and accuracy\n",
    "print('Best alpha: ', best_alpha, '  |  Best Score: ', round(accs.max()*100, 2))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4563e714",
   "metadata": {},
   "source": [
    "Visualizing the Model Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "4bdb1cee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot accuracies per alpha values\n",
    "plt.figure(figsize=(15,7))\n",
    "plt.plot(alphas, accs)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "3bf87fe5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Retrain best performing model\n",
    "best_model = MultinomialNB(alpha=best_alpha)\n",
    "best_model.fit(X_train_tfidf, y_train)\n",
    "\n",
    "# Predict test data with best model\n",
    "probs = best_model.predict_proba(X_test_tfidf)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "2305ba37",
   "metadata": {},
   "source": [
    "Accuracy of Best Model on Sentiment Analysis of tweets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "127d3262",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Naive-Bayes Accuracy: 66.78 %\n"
     ]
    }
   ],
   "source": [
    "# Print accuracy of best performing model on tweet sentiment analysis \n",
    "print('Naive-Bayes Accuracy:', round(len(np.where(y_test == probs.argmax(axis=1))[0])/len(probs) * 100, 2), '%')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4b2c1347",
   "metadata": {},
   "source": [
    "Visualizing the Accuracy of Naive Bayes Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "8ba4dec1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Calculate accuracy\n",
    "accuracy = round(len(np.where(y_test == probs.argmax(axis=1))[0]) / len(probs) * 100, 2)\n",
    "\n",
    "# Create a bar plot\n",
    "plt.figure(figsize=(6, 4))\n",
    "plt.bar(['Naive-Bayes'], [accuracy])\n",
    "plt.xlabel('Model')\n",
    "plt.ylabel('Accuracy (%)')\n",
    "plt.title('Accuracy of Naive-Bayes Model')\n",
    "plt.ylim(0, 100)\n",
    "plt.show()\n"
   ]
  }
 ],
 "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": "3.10.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}