abairedd_HW5 (1).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <Center> Homework 5: Classification </center>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This assignment needs the following two data files:\n",
    "- train.csv: dataset for training\n",
    "- test.csv: dataset for testing. \n",
    "    \n",
    "Both of them have samples in the following format. The `text` column contains documents and the `label` column gives the sentiment of each document.\n",
    "\n",
    "|label | text |\n",
    "|------|------|\n",
    "|1|  I must admit that I'm addicted to \"Version 2.0...|\n",
    "|0|  I think it's such a shame that an enormous tal...|\n",
    "|1|  The Sunsout No Room at The Inn Puzzle has oddl...|\n",
    "|...|...|\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Q1 Classification\n",
    "\n",
    "- Define a function `create_model(train_docs, train_y, test_docs, test_y, tokenizer=None, model_type='svm', stop_words='english', min_df = 1, print_result = True, algorithm_para=1.0`), where\n",
    "\n",
    "    - `train_docs`: is a list of documents for training\n",
    "    - `train_y`: is the ground-truth labels of training documents\n",
    "    - `test_docs`: is a list of documents for test\n",
    "    - `test_y`: is the ground-truth labels of test documents\n",
    "    - `tokenizer`: is the tokenizer and it can be customized by using the lemmatizer of Spacy package\n",
    "    - `model_type`: two options: `nb` (Multinomial Naive Bayes) or `svm` (Linear SVM)\n",
    "    - `stop_words`: indicate whether stop words should be removed. The default value is 'english', i.e. remove English stopwords.\n",
    "    - `min_df`: only word with document frequency above this threshold can be included. The default is 1. \n",
    "    - `print_result`: controls whether to show classification report or plots. The default is True.\n",
    "    - `algorithm_para`: hyperparameter of algorithms, refers to C in SVM and alpha in NB. Both C and alpha have default values of 1.0. Please check sklearn documentation here: https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html\n",
    "    https://scikit-learn.org/stable/modules/generated/sklearn.svm.LinearSVC.html\n",
    "\n",
    "\n",
    "- This function does the following:\n",
    "    - Fit a `TfidfVectorizer` using `train_docs` with options `stop_words, min_df` as specified in the function inputs. Extract features from `train_docs` using the fitted `TfidfVectorizer`.\n",
    "    - Build `linear SVM` or `Multinomial Naive Bayes` model as specified by `model_type` with `algorithm_para` setting. Train the model using the extracted features and `train_y`. \n",
    "    - Transform `test_docs` by the fitted `TfidfVectorizer` (hint: use function `transform` not `fit_transform`).\n",
    "    - Predict the labels for `test_docs`. If `print_result` is True, print the classification report.\n",
    "    - Calculate the AUC score and PRC score (or Average Precision) for class 1 on the test dataset. If `print_result` is True, plot the ROC and PRC curves. **Hint**: \n",
    "        - `sklearn.svm.LinearSVM` does not provide `predict_proba` function. \n",
    "        - Instead, you can use its `decision_function` (see <a href = \"https://stackoverflow.com/questions/59227176/how-to-plot-roc-and-calculate-auc-for-binary-classifier-with-no-probabilities-s\">some reference code</a>) \n",
    "        - Another option is to use `sklearn.svm.SVC` with `kernel='linear' and probability=False` (see <a href = \"https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html\"> reference</a>)\n",
    "    - Return the AUC and PRC scores.\n",
    "\n",
    "    \n",
    "- Test your function with following cases:\n",
    "    - model_type='svm', stop_words = 'english', min_df = 1\n",
    "    - model_type='nb', stop_words = 'english', min_df = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn import svm\n",
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "from sklearn.metrics import roc_curve, auc,precision_recall_curve, average_precision_score\n",
    "from sklearn.metrics import classification_report\n",
    "from matplotlib import pyplot as plt\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.svm import SVC\n",
    "import numpy as np\n",
    "\n",
    "train = pd.read_csv(\"train5.csv\")\n",
    "test = pd.read_csv(\"test5.csv\")\n",
    "# get portion to reduce processing time\n",
    "train=train.iloc[0:5000]\n",
    "test=test.iloc[0:5000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "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>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>17583</td>\n",
       "      <td>1</td>\n",
       "      <td>This makes dull hair look fresh. It's not too ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>19745</td>\n",
       "      <td>0</td>\n",
       "      <td>This book took almost two weeks to receive. No...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>13078</td>\n",
       "      <td>1</td>\n",
       "      <td>As a person who frequents the golf course on a...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9954</td>\n",
       "      <td>0</td>\n",
       "      <td>This is a very needed message for the children...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>16217</td>\n",
       "      <td>1</td>\n",
       "      <td>I purchased the Intex airbed to replace an Aer...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  label                                               text\n",
       "0       17583      1  This makes dull hair look fresh. It's not too ...\n",
       "1       19745      0  This book took almost two weeks to receive. No...\n",
       "2       13078      1  As a person who frequents the golf course on a...\n",
       "3        9954      0  This is a very needed message for the children...\n",
       "4       16217      1  I purchased the Intex airbed to replace an Aer..."
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['run', 'in', 'the', 'park', 'and', 'then', 'run', 'away', 'again', 'fall', 'down']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['run',\n",
       " 'in',\n",
       " 'the',\n",
       " 'park',\n",
       " 'and',\n",
       " 'then',\n",
       " 'run',\n",
       " 'away',\n",
       " 'again',\n",
       " 'fall',\n",
       " 'down']"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# define the customized tokenizer using spacy package\n",
    "# you can also use stanza package\n",
    "# this function tokenize the text and get lemmas\n",
    "\n",
    "import nltk\n",
    "import spacy\n",
    "\n",
    "nlp = spacy.load(\"en_core_web_sm\")\n",
    "\n",
    "def MyTokenizer(text):\n",
    "    doc=nlp(text)\n",
    "    newTokens = []\n",
    "    \n",
    "    for word in doc:\n",
    "\n",
    "        res = word.lemma_\n",
    "        newTokens.append(res)\n",
    "    print(newTokens)\n",
    "    return newTokens\n",
    "\n",
    "MyTokenizer('Runing in the park and then ran away again fell down')\n",
    "        \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the function for Task 1\n",
    "\n",
    "def create_model(train_docs, train_y, test_docs, test_y, tokenizer=None, model_type=\"svm\", min_df = 1,stop_words=None, print_result = True, algorithm_para=1.0):\n",
    "    \n",
    "    # initialize\n",
    "    tfidf_vect = TfidfVectorizer(stop_words=stop_words,min_df=min_df)\n",
    "    \n",
    "    # generate \n",
    "    dtm=tfidf_vect.fit_transform(train_docs)\n",
    "    \n",
    "    if(model_type=='svm'):\n",
    "        clf= svm.LinearSVC(C=algorithm_para)\n",
    "    else:\n",
    "        clf= MultinomialNB(alpha=algorithm_para)\n",
    "        \n",
    "    clf.fit(dtm,train_y)\n",
    "    \n",
    "    # transform \n",
    "    dtm_test=tfidf_vect.transform(test_docs)\n",
    "    \n",
    "    predicted=clf.predict(dtm_test)\n",
    "    \n",
    "    # calculate \n",
    "    cf=classification_report(test_y,predicted)\n",
    "    \n",
    "   # auc\n",
    "    binary_y = np.where(test_y,1,0)\n",
    "    \n",
    "    if(model_type=='svm'):\n",
    "        predict_p=clf.decision_function(dtm_test)\n",
    "        fpr, tpr, thresholds = roc_curve(binary_y, predict_p, pos_label=1)\n",
    "    else:\n",
    "        predict_p=clf.predict_proba(dtm_test)\n",
    "        fpr, tpr, thresholds = roc_curve(binary_y, predict_p[:,1], pos_label=1)\n",
    "    \n",
    "    # prc\n",
    "    if(model_type=='svm'):\n",
    "        precision, recall, thresholds = precision_recall_curve(binary_y, predict_p, pos_label=1)\n",
    "    else:\n",
    "        precision, recall, thresholds = precision_recall_curve(binary_y, predict_p[:,1], pos_label=1)\n",
    "        \n",
    "    if(print_result):\n",
    "        print(cf)\n",
    "        \n",
    "    auc_score=auc(fpr, tpr)\n",
    "    \n",
    "    if(model_type=='svm'):\n",
    "        prc_score=average_precision_score(binary_y,predict_p)\n",
    "    else:\n",
    "        prc_score=average_precision_score(binary_y,predict_p[:,1])\n",
    "        \n",
    "    if(print_result):\n",
    "        print(\"AUC: {:.2%}\".format(auc_score),\"PRC: {:.2%}\".format(prc_score))\n",
    "    \n",
    "    # line charts \n",
    "        if(model_type=='svm'):\n",
    "            plt.figure();\n",
    "            plt.plot(fpr, tpr, color='darkorange', lw=2);\n",
    "            plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--');\n",
    "            plt.xlim([0.0, 1.0]);\n",
    "            plt.ylim([0.0, 1.05]);\n",
    "            plt.xlabel('False Positive Rate');\n",
    "            plt.ylabel('True Positive Rate');\n",
    "            plt.title('svm-AUC');\n",
    "            plt.show();\n",
    "    \n",
    "            plt.figure();\n",
    "            plt.plot(recall, precision, color='darkorange', lw=2);\n",
    "            plt.xlim([0.0, 1.0]);\n",
    "            plt.ylim([0.0, 1.05]);\n",
    "            plt.xlabel('Recall');\n",
    "            plt.ylabel('Precision');\n",
    "            plt.title('svm-PRC');\n",
    "            plt.show();\n",
    "        else:\n",
    "            plt.figure();\n",
    "            plt.plot(fpr, tpr, color='darkorange', lw=2);\n",
    "            plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--');\n",
    "            plt.xlim([0.0, 1.0]);\n",
    "            plt.ylim([0.0, 1.05]);\n",
    "            plt.xlabel('False Positive Rate');\n",
    "            plt.ylabel('True Positive Rate');\n",
    "            plt.title('nb-AUC');\n",
    "            plt.show();\n",
    "\n",
    "            plt.figure();\n",
    "            plt.plot(recall, precision, color='darkorange', lw=2);\n",
    "            plt.xlim([0.0, 1.0]);\n",
    "            plt.ylim([0.0, 1.05]);\n",
    "            plt.xlabel('Recall');\n",
    "            plt.ylabel('Precision');\n",
    "            plt.title('nb-PRC'); \n",
    "            plt.show();\n",
    "\n",
    "    return auc_score, prc_score\n",
    "    \n",
    "\n",
    "    \n",
    "    \n",
    "\n",
    "   \n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.83      0.81      0.82       509\n",
      "           1       0.80      0.83      0.82       491\n",
      "\n",
      "    accuracy                           0.82      1000\n",
      "   macro avg       0.82      0.82      0.82      1000\n",
      "weighted avg       0.82      0.82      0.82      1000\n",
      "\n",
      "AUC: 90.08% PRC: 89.92%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "auc_score, prc_socre = create_model(train[\"text\"], train[\"label\"], test[\"text\"], test[\"label\"], model_type='svm',  min_df = 1, stop_words='english',   print_result=True, algorithm_para=1.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.84      0.84      0.84       509\n",
      "           1       0.84      0.84      0.84       491\n",
      "\n",
      "    accuracy                           0.84      1000\n",
      "   macro avg       0.84      0.84      0.84      1000\n",
      "weighted avg       0.84      0.84      0.84      1000\n",
      "\n",
      "AUC: 90.67% PRC: 90.97%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.84      0.84      0.84       509\n",
      "           1       0.84      0.84      0.84       491\n",
      "\n",
      "    accuracy                           0.84      1000\n",
      "   macro avg       0.84      0.84      0.84      1000\n",
      "weighted avg       0.84      0.84      0.84      1000\n",
      "\n",
      "AUC: 90.67% PRC: 90.97%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Test the function with 'svm'\n",
    "# Compare the default tokenizer and customized tokenizer\n",
    "# Explain your finding\n",
    "\n",
    "auc_score, prc_socre = create_model(train[\"text\"], train[\"label\"], test[\"text\"], test[\"label\"], \\\n",
    "            tokenizer=None,model_type='svm',  min_df = 1, stop_words=None, \\\n",
    "                                    print_result=True, algorithm_para=1.0)\n",
    "\n",
    "auc_score, prc_socre = create_model(train[\"text\"], train[\"label\"], test[\"text\"], test[\"label\"], \\\n",
    "            tokenizer=MyTokenizer,model_type='svm', min_df = 1, stop_words=None, \\\n",
    "                                    print_result=True, algorithm_para=1.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.82      0.81       509\n",
      "           1       0.81      0.80      0.80       491\n",
      "\n",
      "    accuracy                           0.81      1000\n",
      "   macro avg       0.81      0.81      0.81      1000\n",
      "weighted avg       0.81      0.81      0.81      1000\n",
      "\n",
      "AUC: 88.55% PRC: 88.69%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.82      0.81       509\n",
      "           1       0.81      0.80      0.80       491\n",
      "\n",
      "    accuracy                           0.81      1000\n",
      "   macro avg       0.81      0.81      0.81      1000\n",
      "weighted avg       0.81      0.81      0.81      1000\n",
      "\n",
      "AUC: 88.55% PRC: 88.69%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Test the function with 'nb' \n",
    "# Compare the default tokenizer and customized tokenizer\n",
    "# Explain your finding\n",
    "\n",
    "auc_score, prc_socre = create_model(train[\"text\"], train[\"label\"], test[\"text\"], test[\"label\"], \\\n",
    "            tokenizer=MyTokenizer,model_type='nb',  min_df = 1, stop_words=None, \\\n",
    "                                    print_result=True, algorithm_para=1.0)\n",
    "\n",
    "\n",
    "\n",
    "auc_score, prc_socre = create_model(train[\"text\"], train[\"label\"], test[\"text\"], test[\"label\"], \\\n",
    "            tokenizer=None,model_type='nb', min_df = 1, stop_words=None, \\\n",
    "                                    print_result=True, algorithm_para=1.0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Task 2: Search for best parameters \n",
    "\n",
    "From Task 1, you may find there are many possible ways to configure parameters. Next, let's use grid search to find the optimal parameters\n",
    "\n",
    "- Define a function `search_para(docs, y)` where `docs` are training documents and `y` is the ground-truth labels.\n",
    "- This function does the following:\n",
    "    - Create a pipleline which integrates `TfidfVectorizer` and `SVM` classifier \n",
    "    - Define the parameter ranges as follow:\n",
    "        - `stop_words': [None, 'english']`\n",
    "        - `min_df: [1,2,5]`\n",
    "        - `C: [0.1,0.5,1]`\n",
    "    - Set the scoring metric to \"f1_macro\"\n",
    "    - Use `GridSearchCV` with `5-fold cross validation` to find the best parameter values based on the training dataset. \n",
    "    - Print the best parameter values\n",
    "    \n",
    "- Call the function `svm_model` defined in Task 1 (default tokenizer) `with the best parameter values`. You will have to add the best C value to the original function, because you used default settings of the SVM classifier previously. Then please briefly answer the following:\n",
    "    - Compare with the model in Task 1, how is the performance improved on the test dataset?\n",
    "    - Why do you think the new parameter values help sentiment classification?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define grade search function\n",
    "\n",
    "def search_para(docs, y):\n",
    "    text_clf = Pipeline([('tfidf', TfidfVectorizer()),\n",
    "                     ('clf', svm.LinearSVC())\n",
    "                   ])\n",
    "    \n",
    "    parameters = {'tfidf__min_df':[1, 2,5],\n",
    "              'tfidf__stop_words':[None,\"english\"],\n",
    "              'clf__C': [0.1, 0.5,1],}\n",
    "    \n",
    "    metric =  \"f1_macro\"\n",
    "    \n",
    "    gs_clf = GridSearchCV(text_clf, param_grid=parameters, scoring=metric, cv=5, n_jobs=-1)\n",
    "    gs_clf = gs_clf.fit(docs,y)\n",
    "\n",
    "    return gs_clf.best_params_\n",
    "\n",
    "    \n",
    "    \n",
    "  \n",
    "\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'clf__C': 0.5, 'tfidf__min_df': 1, 'tfidf__stop_words': None}"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "search_para(train[\"text\"], train[\"label\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.84      0.85       509\n",
      "           1       0.84      0.85      0.84       491\n",
      "\n",
      "    accuracy                           0.84      1000\n",
      "   macro avg       0.84      0.84      0.84      1000\n",
      "weighted avg       0.84      0.84      0.84      1000\n",
      "\n",
      "AUC: 90.91% PRC: 91.29%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEWCAYAAAB42tAoAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAwYElEQVR4nO3dd5hU5fn/8ffNLrt0EBYLKAKKgAWiLmJDUCBUo8bEGozmm8iCYFf8YddYsIAoTYJGEwuJSpQAimIBgkpTmjQpiiigSJG6bLl/f8yAk80yzC47c2Z2P6/r2mvnzJxz5t4jzmee85zzPObuiIiI7E+loAsQEZHkpqAQEZGoFBQiIhKVgkJERKJSUIiISFQKChERiUpBISIiUSkoROLIzK42MzezS4p5/j/FrP+VmXWKWD7NzCaZ2RYz22Rms8zsmkTULrKXgkIkvn4PbAr/LhEzOwP4AJgKHAvUA/oA3cqyQJEDUVBIhWBmA8zsWzPbZmbLzOxKM9tlZnUj1jnZzDaaWeXwN/4ZZjYk/G1+lZmdGX7+GzP73syifvib2dFAe+BaoIuZHVbCsh8HXnT3Qe6+0UPmuvslB9xSpAwpKKTcM7PmQD+gjbvXBLoAnwKfABdHrHoF8Lq754WX2wILCH2TfwUYC7Qh9O3+d8AwM6sR5a2vAua4+xvAEuDKEtRcDTgDeD3WbUTiRUEhFUEBkAkcb2aV3f0rd19J6MP/cgAzM+Cy8HN7rXb3v7p7AfAP4CjgAXfPdfd3gT2EQmN/rorY3yuU7PTTIYT+/1xXgm1E4kJBIeWeu68AbgTuA743s7Fm1oDQt/Uzwo/PARyYHrHphojHu8L7KvpcDTNrZGbb9/4AmNlZQBNCrRAIBcVJZvaL8HI+ULmYcisDecBmoBA4ojR/s0hZUlBIheDur7j72cDRhAJhkLtvAd4FLiF02ulVL8Vwyu6+xt1r7P0JP/17wIB5ZrYemBl+/qrw7zVAo3BLBth3uulQ4Gt338n/nhoTCYSCQso9M2tuZueZWSawm1BLoCD88iuEPrwv5r9POx3M+1UhFD7XAr+I+OkPXGlm6YSCYzdwh5lVMbPqwKPAHODr8K5uB642s9vMrF54363NbCwiCaSgkIogk9CH8EZgPaFv7QPDr40HmgEb3H1+Gb3fhYTC6G/uvn7vD/AckAZ0dfdcoAfQAVgLrAIaAJfsbdW4+8fAeeGfVWa2CRgNTCqjOkViYpq4SEREolGLQkREolJQiIhIVAoKERGJSkEhIiJRpQddQEllZWV548aNgy5DRCSlzJ07d6O71y/NtikXFI0bN2bOnDlBlyEiklLM7OsDr1U8nXoSEZGoFBQiIhKVgkJERKJSUIiISFQKChERiUpBISIiUcUtKMzs+fC8wov287qZ2dNmtsLMFpjZKfGqRURESi+eLYoXgK5RXu9GaHjnZoTG7R8Zx1pERCqsPXsKDrxSFHG74c7dp5lZ4yirXEBovH4HPjWzOmZ2hLtrjmARKZ/G9YDViZ1OZOj0toyZeXAnbIK8M7sh8E3E8trwc/8TFGZ2LaFWB40aNUpIcSKSAAF8cFY0rY/YwOINpRq5Y58gg8KKea7YWZTcfTShmb3Izs7WTEsiqUAhULwm3eHXE+O2+2++2cqECcvp06cNEJpCccXtm2na9IFS7zPIoFgLHBWxfCTwXUC1iEg08fzQj/MHZ0WRn1/I00/P5J57PmTHjjxOPPFQ2rU7GoAmTQ45qH0HGRTjgX7hieLbAlvVPyGSQIn4xq8QSIiZM9fSu/cE5s/fAMDFF7ekadODC4dIcQsKM3uVUKsny8zWAvcClQHcfRShCeK7AyuAncA18apFpEIq6yDQh37S2bx5FwMHvs+zz87FHRo3rsOwYd3o0eO4Mn2feF71dPkBXnfguni9v0iFcTCBoA//lHb//VMZNWou6emVuPXWM7j77vZUq1a5zN8n5eajEKmQShsGCoJyJz+/kPT00C1wd911DqtXb+Ghh87jxBMPjdt7KihEkoVaBhLF7t35DBr0H958cxkzZ/6RjIw0srKq8dZbl8X9vRUUIvGiPgIpI++/v4o+fSby5ZebAJg8eQXnn988Ye+voBApiXhfKaQwkAgbNmznllve5eWXFwLQsmUWI0f2oH37xgmtQ0EhEs3BBoM++KWUXnppAf37v82WLbupUiWde+45h1tuOZOMjLSE16KgECnqQOGgD39JgMJCZ8uW3XTteizDh3cv0/siSkpBIRVTSVoKCgZJgO3b9/DJJ9/QufMxAPTq1YoGDWrSsWMTzIob8ShxFBSS+uLRb6BwkAR6882l9O//Nj/8sINFi/py7LF1MTM6dWoadGmAgkJShcJAyqGvv97C9de/w/jxywDIzm5Abm5+wFX9LwWFJBfdWCYVQF5eAU899Sn33TeVnTvzqFkzg4cf7kifPtmkpSXfDNUKCkk8hYFUcNdf/zajRs0F4JJLTmDIkC40aFAz4Kr2T0Eh8VWaUFAgSDl3442nM3Xq1wwe3IWuXY8NupwDUlDIwVHrQCQqd+ellxYwadIKXnnl15gZzZtnsWhRXypVCvZqplgpKKTkShoOCgWpoJYt20ifPhP58MOvgNAlr927NwNImZAABYVEUutApEzs2pXHI4/8h0GDZrBnTwH16lXlySd/SbduyX+aqTgKiopCA9SJJMSUKavIyZnAypWbAfi//zuZQYM6Ua9etYArKz0FRXml1oFIID7++BtWrtzMCSfUZ9Sonpx9dqOgSzpoCoryJlpAKAREylxBQSErVmyiefMsAAYMOIusrGr88Y+nBDKAXzwoKMqL4gJCwSASV59/vo6cnImsWrWZZcv6UbduVTIz0+nbt03QpZUpBUV5ERkSCgiRuNq2LZd77vmQp5+eRWGh07BhTVau3ETdug2DLi0uFBSprmhL4hYPrhaRcs7dGTduCTfc8A7ffruNSpWMm246nfvv70DNmplBlxc3CopUc6A+CBGJmxtvfIenn54FQJs2DXj22Z6cfPIRAVcVfwqKVKCJdESSwkUXteTFF+fz8MMd6d371KQcwC8eFBTJbH8BoWAQSYj//GcNH364mrvvbg9Ahw6NWbPmJmrVKr+nmYqjoAharPc7KBxEEubHH3cyYMAUnnvucwA6dmzKmWceBVDhQgIUFMGKJSQUECIJ4+787W/zufXW99i4cSeVK1fijjvO5uSTDw+6tEApKIJQNCAUBiKBW7LkB/r0mcjUqV8DcO65jRkxogctWmQFXFnwFBSJpJviRJLW4MGfMHXq19SvX43Bg7tw5ZUnYZY6I7zGk4IiUdSKEEk6W7fupnbtKgA88kgnqlfP4J572lO3btWAK0suCopE2RsSCgiRwH333TZuumkyCxZsYP78HDIy0sjKqsZTT3UNurSkpKCIp+JONSkkRAJTUFDIiBGzufPOD9i2bQ/VqlXms8/WcfrpRwZdWlJTUMRTcf0RIhKIuXO/o3fvCcyduw6AX/2qOc88041GjWoHXFnyi2tQmFlXYCiQBoxx90eLvF4beAloFK7lCXf/azxrCoTGXxIJ1H33fcSDD06jsNA56qhaPPNMNy64oEXQZaWMuAWFmaUBw4HOwFpgtpmNd/fFEatdByx29/PNrD6wzMxedvc98apLRCqepk0PwQxuueUM7ruvAzVqZARdUkqJZ4viNGCFu68CMLOxwAVAZFA4UNNC16DVADYB+XGsKf7KespRESmxVas2M3v2t1x66YkA9OrVirZtG+6bXEhKJp5B0RD4JmJ5LdC2yDrDgPHAd0BN4FJ3Lyy6IzO7FrgWoFGjJJ5WcH/3SYhIQuzZU8ATT3zMgw9Ow9059dQGHHtsXcxMIXEQ4hkUxd2pUvRkfRdgHnAecAzwnplNd/ef/msj99HAaIDs7OzkPeGvS2BFAjNt2tfk5ExgyZKNAFx55UkVclymeIhnUKwFjopYPpJQyyHSNcCj7u7ACjNbDbQAZsWxrrJXtCWhkBBJmI0bd3Lbbe/xwgvzAGjWrC4jR/agY8emwRZWjsQzKGYDzcysCfAtcBlwRZF11gAdgelmdhjQHFgVx5rKlk41iQQuJ2cCb7yxhMzMNAYObMftt59FlSq68r8sxe1ounu+mfUDJhO6PPZ5d//CzHLCr48CHgReMLOFhE5VDXD3jfGqqcxpSA6RQBQWOpUqhc5uP/TQeezalc9TT3WhWbN6AVdWPlnorE/qyM7O9jlz5gRbhOapFgnEzp15PPjgVObN28CkSVdo0L4SMLO57p5dmm3VPisJnWoSCczEicvp1+9tvvpqC2Ywa9a3tG2roTcSQUERCw0PLhKYtWt/4oYb3mHcuCUAtG59GKNG9VRIJJCCIhbqixAJxIgRsxkwYArbt++hevXKPPjgufTv35b09EpBl1ahKChKQn0RIgm1ceNOtm/fw0UXtWDo0K4cdZQG8AuCgkJEksaWLbtZunTjvmG/Bww4i9NOa0jXrscGXFnFpvbbgYzrEXQFIuWeuzN27CJathzOr371Kps27QIgMzNdIZEEFBTRRHZi6+omkbhYsWITXbu+zOWXv8H69dtp1qweW7fuDrosiaBTT/tTNCTUgS1SpnJz83nssRk89NB0cnMLOOSQKjz2WGf+8IeT991MJ8kh5qAws+ruviOexSQVhYRIXF166eu89dYyAK66qjWPP96ZQw+tHnBVUpwDBoWZnQmMITRfRCMzaw30dve+8S4uofY3j4RCQiQubrzxdJYt+5ERI7pz7rlNgi5HooilRTGE0HDg4wHcfb6ZnRPXqhIp2kRD6pcQKROFhc7zz3/OkiU/8OSTXQDo0KExixb1IS1NXaXJLqZTT+7+TZExVQriU04AdDOdSFwtXLiBnJyJfPxxaB6zq65qTevWhwMoJFJELEHxTfj0k5tZBnA9sCS+ZQVAN9OJlKkdO/Zw//1TGTz4EwoKnMMPr8FTT3WhVavDgi5NSiiWoMgBhhKa2nQt8C5QvvonRKRM/fvfy+jX723WrNmKGVx3XRseeug8ateuEnRpUgqxBEVzd78y8gkzOwuYEZ+SRCTVvfnmUtas2crJJx/Os8/2pE2bhkGXJAchlqB4BjglhudSS7RObBEpkfz8Qr799ieOProOAIMGdebkk48gJydbA/iVA/sNCjM7AzgTqG9mN0e8VIvQjHWprWgntoiUyqefriUnZwK5uQXMn59DRkYaWVnV6NfvtKBLkzISrUWRQejeiXSgZsTzPwG/iWdRCaVObJFS2bx5FwMHvs+zz87FHRo3rsNXX23huOM0HWl5s9+gcPepwFQze8Hdv05gTSKSxNydV19dxE03Teb773eQnl6J2247k7vuOodq1SoHXZ7EQSx9FDvN7HHgBGDfJQvufl7cqhKRpHXlleN49dVFALRr14iRI3twwgmHBlyVxFMsvUwvA0uBJsD9wFfA7DjWFH8aOlyk1Lp2PZZ69ary/PO/4qOPrlZIVACxtCjquftzZnZDxOmoqfEuLK40dLhIzKZMWcXKlZvo3TsbgF69WtGz53HUrVs14MokUWIJirzw73Vm1gP4Digfs5pruA6R/dqwYTs33/wur7yykMzMNDp1asoxx9TFzBQSFUwsQfFnM6sN3ELo/olawI3xLEpEglNY6IwePZc77pjC1q25VKmSzj33nKP5qiuwAwaFu08IP9wKnAv77swWkXJm/vz19O49gZkzvwWgW7djGTasO02bHhJwZRKkaDfcpQGXEBrj6R13X2RmPYGBQFXg5MSUWMbUkS2yX7ffPoWZM7+lQYOaDB3alYsvbkmRkaOlAorWongOOAqYBTxtZl8DZwB3uPubCagtPtSRLbKPu7NzZx7Vq2cA8PTTXRk1ag73338utWplBlydJItoQZENtHL3QjOrAmwEjnX39YkpLc7UkS0V3Ndfb6F//7fZsSOPKVN6YWY0b57FkCFdgy5Nkky0oNjj7oUA7r7bzJaXm5AQqcDy8goYMuRT7r9/Kjt35lGzZgZffrlJQ2/IfkULihZmtiD82IBjwssGuLu3int1IlKmZsxYQ07ORBYt+h6ASy89gcGDu9CgQc0DbCkVWbSgaJmwKhJFHdlSgfXvP4lhw0KDKjRtegjDh3ena9djA65KUkG0QQHL30CA6siWCqx+/epUrlyJAQPOYuDAdlStqgH8JDZxnVHEzLqa2TIzW2Fmd+xnnQ5mNs/Mvojr0CCRrQl1ZEsFsHTpRt59d+W+5QEDzmLBgj48+OB5CgkpkVjuzC6V8H0Yw4HOhObanm1m4919ccQ6dYARQFd3X2Nm8RtdTK0JqSB27crj4YenM2jQDOrUqcLSpf2oW7cqmZnptGiRFXR5koJiCgozqwo0cvdlJdj3acAKd18V3sdY4AJgccQ6VwDj3H0NgLt/X4L9l45aE1KOvfvuSvr2ncjKlZsB+NWvmqP75eRgHfDUk5mdD8wD3gkv/8LMxsew74bANxHLa8PPRToOOMTMPjKzuWZ2VUxVi8h/WbduG5dd9jpdurzEypWbOeGE+kyffg1jxvyKQw7RAH5ycGJpUdxHqHXwEYC7zzOzxjFsV9z3mKLzjqYDpwIdCQ0L8omZferuy/9rR2bXAtcCNGrUKIa3jjCux3/Pjy1SDv361//k00/XUrVqOvfd14GbbjqdypVTf2p7SQ6xdGbnu/vWUux7LaEhQPY6ktAQ5UXXecfdd7j7RmAa0Lrojtx9tLtnu3t2/fr1S1ZFZEiof0LKEfefv3c9+mhHevY8jsWLr+P2289SSEiZiqVFscjMrgDSzKwZcD3wcQzbzQaamVkT4FvgMkJ9EpHeAoaZWTqQAbQFhsRafIncUrQxI5Katm3L5Z57PmTHjjxGjz4fgPbtG9O+feNgC5NyK5YWRX9C82XnAq8QGm78xgNt5O75QD9gMrAE+Ke7f2FmOWaWE15nCaG+jwWEBh8c4+6LSvF3iJR77s4bbyymZcvhPPXUTP7613l89dWWoMuSCiCWFkVzd78TuLOkO3f3ScCkIs+NKrL8OPB4SfctUpGsXr2Zfv3eZtKkLwE47bSGjBrVg8aN6wRbmFQIsQTFYDM7AngNGOvuX8S5prKjITskxbk7jz02g/vvn8quXfnUrp3JI4905NprTyUtLa73y4rsE8sMd+ea2eGEJjEabWa1gH+4+5/jXt3B0k12kuLMjOXLf2TXrnwuv/xEBg/uwuGH1wi6LKlgLPLKiQOubHYScDtwqbtnxK2qKLKzs33OnDmxrfxk+ApddWRLCtm4cSfr12/nxBMP3bf8+efr6Nz5mIArk1RmZnPdPbs028Zyw11LM7vPzBYBwwhd8XRkad4soXTaSVKMu/PCC/No0WIYv/3ta+zZUwBAVlY1hYQEKpY+ir8CrwK/dPei90EkL512khSyZMkP5ORMZNq00KDNrVsfzubNuzjsMJ1mkuDF0kdxeiIKKVMaKVZSxM6deTz00DQef/xj8vIKqV+/GoMHd+HKK0/CNEiTJIn9BoWZ/dPdLzGzhfz30BvJPcNd5JAdak1IEnN3zjvvRWbO/BaA3r1P5ZFHOmpsJkk60VoUN4R/90xEIWUmMiTUmpAkZmb07duGnTvzePbZnpxxxlEH3kgkAPvtzHb3deGHfd3968gfoG9iyjsICglJMgUFhTzzzEwGD/5k33O9erVi7txrFRKS1GK5Y6dzMc91K+tCyoSudJIkNWfOd7RtO4brr3+HgQPf57vvtgGhVoUG8JNkF62Pog+hlkNTM1sQ8VJNYEa8CysV9U1Iktm6dTd33fUBw4fPxh2OOqoWzzzTjQYNagZdmkjMovVRvAK8DTwCRM53vc3dN8W1qoOl004SMHfntdcWc+ON77Bu3XbS0oybbjqde+/tQI0agdyrKlJq0YLC3f0rM7uu6AtmVjfpw0IkYM8+O5d167Zz+ulHMmpUD1q3PjzokkRK5UAtip7AXEKXx0Ze1O1A0zjWJZJycnPz2bJlN4cdVgMzY8SI7nz00Vf86U+nUqmS7omQ1LXfoHD3nuHfTRJXjkhqmjr1K3JyJtKgQU2mTOmFmdG8eRbNm2cFXZrIQYtlrKezzKx6+PHvzGywmZVw4uoE0BVPEoAfftjB1Ve/SYcOL7J06Ua++WYrGzbsCLoskTIVy+WxI4GdZtaa0MixXwN/j2tVpaErniSBCgud5577jBYthvPii/PJzEzj/vs7sGBBHw0DLuVOLIMC5ru7m9kFwFB3f87Mfh/vwkpNVzxJnLk7Xbq8xJQpqwDo1KkpI0Z0p1mzegFXJhIfsQTFNjP7f0AvoJ2ZpQGV41uWSPIyM9q1a8TChRsYMqQLl112ogbwk3ItllNPlwK5wB/cfT3QEM1xLRXMxInLefPNpfuWBww4i6VL+3H55RrlVcq/WIYZX29mLwNtzKwnMMvd/xb/0mIUOVqsSBlbu/YnbrjhHcaNW0JWVjXOOedo6tatSmZmOpmZsTTIRVJfLFc9XQLMAn5LaN7smWb2m3gXFrPIkFBHtpSR/PxChgz5hJYthzNu3BKqV6/MwIFnU6tWZtCliSRcLF+J7gTauPv3AGZWH5gCvB7PwkpM82JLGZk161t6957AvHnrAbjoohYMHdqVo46qHXBlIsGIJSgq7Q2JsB+JrW9DJOUUFjrXXPMWixf/QKNGtRk2rBvnn9886LJEAhVLULxjZpMJzZsNoc5tdQpIueHu5OYWUKVKOpUqGcOHd+ftt7/knnvaU726BvATiaUz+zYz+zVwNqHxnka7+7/iXplIAqxYsYm+fSdy1FG1eO65CwDo0KExHTo0DrYwkSQSbT6KZsATwDHAQuBWd/82UYWJxFNubj6DBs3g4Yenk5tbQN26VXnssZ3Uq1ct6NJEkk60vobngQnAxYRGkH0mIRWJxNkHH6ymVatR3HvvR+TmFvD737dm6dLrFBIi+xHt1FNNd/9L+PEyM/ssEQWJxEtBQSHXXPMWf/97aMLG5s3rMWpUT51mEjmAaEFRxcxO5ud5KKpGLru7gkNSSlpaJdLTK1GlSjp33dWOW289UzfNicTA3Iu//8DMPoyynbv7efEpKbrs7GyfM2dOaCHyrmzdRyHFWLhwA7t359OmTUMAfvxxJ1u27OaYY+oGXJlIYpnZXHfPLs220SYuOrf0JSWIhhaX/dixYw/33fcRQ4Z8SrNm9Zg/P4eMjDTq1aumvgiREiof7W4NLS4Rxo9fRv/+b7NmzVbMoFOnJuTlFZCRkRZ0aSIpKa53WJtZVzNbZmYrzOyOKOu1MbOCpBpDSlLOmjVbufDCsVxwwVjWrNnKKaccwaxZf+KZZ7rrxjmRgxC3FkV43orhQGdgLTDbzMa7++Ji1hsETC7RG2jqU4lQUFBIhw4vsHr1FmrWzODPfz6Pvn3bkJ6u0WZEDtYBg8JCg+1fCTR19wfC82Uf7u6zDrDpacAKd18V3s9Y4AJgcZH1+gNvAG1KVLn6J4TQ8BtmRlpaJe67rwP//vdynnqqCw0b1gq6NJFyI5avWyOAM4DLw8vbCLUUDqQh8E3E8trwc/uYWUPgImBUtB2Z2bVmNsfM5vzwww///aL6JyqkzZt3kZMzgYcfnr7vuV69WvHaa79VSIiUsVhOPbV191PM7HMAd99sZrGc8C1u2q+i17A+BQxw94Jos4S5+2hgNIQuj43hvaWccndeeWUhN9/8Lt9/v4OaNTPo1+80ateuopnmROIklqDIC/cjOOybj6Iwhu3WAkdFLB8JfFdknWxgbPh/8Cygu5nlu/ubMexfKpjly3+kb9+JvP/+agDatWvEyJE9qF27SsCViZRvsQTF08C/gEPN7CHgN8BdMWw3G2hmZk2Ab4HLgCsiV3D3Jnsfm9kLwASFhBSVn1/In/88jUce+Q979hRQr15VHn+8M1df/Qu1IkQSIJZhxl82s7lAR0Knky509yUxbJdvZv0IXc2UBjzv7l+YWU749aj9EiJ7paUZ06evYc+eAv7wh18waFBnsrJ005xIoux3CI99K4Sucvof7r4mLhUdwL4hPJ4Mf5PU0B3l0oYN29m9O5+jj64DwJdf/si6dds555yjgy1MJEXFZQiPCBMJ9U8YUAVoAiwDTijNG4pEU1jojB49lzvumEJ2dgPee68XZkazZvVo1qxe0OWJVEixnHo6KXLZzE4BesetIqmw5s1bT07OBGbODM2PlZGRxvbte6hZMzPgykQqthLfme3un5lZyW6OE4li27Zc7r33I4YOnUlhodOgQU2GDu3KxRe3VGe1SBKI5c7smyMWKwGnAD/sZ3WREtmzp4BTThnNihWbqFTJuOGGtjzwwLnUqqVWhEiyiKVFUTPicT6hPos34lOOVDQZGWn06tWKf/97OaNG9eDUUxsEXZKIFBE1KMI32tVw99sSVI+Uc3l5BQwZ8imNGtXmsstOBOCOO87mzjvbkZamAfxEktF+g8LM0sP3QpySyIKk/JoxYw05ORNZtOh76tevRs+ex1GjRobmiRBJctFaFLMI9UfMM7PxwGvAjr0vuvu4ONcm5cSmTbsYMOA9xoz5HICmTQ9hxIju1KihOSJEUkEsfRR1gR+B8/j5fgoHFBQSlbvz978v4JZb3mXjxp1UrlyJAQPOYuDAdlStWjno8kQkRtGC4tDwFU+L+Dkg9tLt0HJAeXmFPPLIf9i4cSft2x/NyJE9aNmyftBliUgJRQuKNKAGsQ0XLgLArl157NlTQO3aVcjISGP06J6sWrWZq65qrXsiRFJUtKBY5+4PJKwSSXmTJ6+gb99JdOhwNM89dwEA7dodTbt2Gp9JJJVFCwp9/ZOYrFu3jZtumsw//vEFANWrV2bnzjyqVVM/hEh5EO3C9Y4Jq0JSUkFBIcOGzaJFi+H84x9fULVqOoMGdWLu3GsVEiLlyH5bFO6+KZGFSGrZvTufc875K7NnhyYt7NnzOJ55phuNG9cJtjARKXMlHhRQBKBKlXROPPFQ1q3bztNPd+XCC1uos1qknErNoBjXI+gKKhx3Z9y4JRx2WA3OPjs0l9XgwV1ISzMNAy5SzqVmUKyeFPrdpHuwdVQQq1dvpl+/t5k06UtatMhi3rzeZGamU6dOlaBLE5EESM2g2OvXE4OuoFzbs6eAJ5/8mAcfnMauXfnUrp3JDTe0JT1dg/eJVCSpHRQSN9Onf01OzkQWLw5NPXLFFSfx5JO/5PDDawRcmYgkmoJC/seuXXn85jev8f33Ozj22LqMGNGdzp2PCbosEQmIgkKAUGd1QYGTnl6JqlUrM3jwL1m+/Ef+3/9rR5Uq+mciUpHpE0BYvPgHcnIm0LlzU+6+uz0AV17ZKuCqRCRZqFeyAtu5M4+BA9+ndetRTJ++hjFjPic3Nz/oskQkyahFUUG9/faXXHfdJFav3gJA796n8sgjHcnM1D8JEflv+lSoYHbs2MPVV7/F668vBqBVq8MYNaoHZ5xxVMCViUiyUlBUMNWqVWbTpl1Ur16Z++/vwA03nK77IkQkKgVFBTBnznfUqVOFY4+ti5kxZsz5pKVVolGj2kGXJiIpQF8ly7GtW3fTv/8kTjvtL+TkTMA9NDFhkyaHKCREJGZqUZRD7s4///kFN944mfXrt5OWZpxyyhHk5xdSuXJa0OWJSIpRUJQzK1du4rrrJjF58koAzjjjSEaN6kmrVocFXJmIpCoFRTmybVsu2dl/YcuW3dSpU4VBgzrxxz+eQqVKmidCREovrkFhZl2BoUAaMMbdHy3y+pXAgPDidqCPu8+PZ03lWc2amdx00+msWLGJJ574JYceWj3okkSkHIhbUJhZGjAc6AysBWab2Xh3Xxyx2mqgvbtvNrNuwGigbbxqKm9++GEHt932Hh07NqFXr9YA3H33OZppTkTKVDyvejoNWOHuq9x9DzAWuCByBXf/2N03hxc/BY6MYz3lRmGhM2bMZzRvPowXX5zPnXd+QF5eAYBCQkTKXDxPPTUEvolYXkv01sL/AW8X94KZXQtcC9CoUaOyqi8lLVr0PTk5E5gxI3RoO3VqyogR3XU1k4jETTyDorivtl7simbnEgqKs4t73d1HEzotRXZ2tsOasqoxZezalcd9933E4MGfkp9fyGGHVWfIkC5cdtmJakWISFzFMyjWApEDCB0JfFd0JTNrBYwBurn7j3GsJ6VVqmSMH7+cgoJC+vbN5qGHOmrOahFJiHgGxWygmZk1Ab4FLgOuiFzBzBoB44Be7r48jrWkpLVrf6JatcrUrVuVzMx0Xngh1MXTtq26ckQkceLWme3u+UA/YDKwBPinu39hZjlmlhNe7R6gHjDCzOaZ2Zx41ZNK8vMLGTLkE1q2HM5tt7277/m2bY9USIhIwsX1Pgp3nwRMKvLcqIjHfwT+GM8aUs3MmWvp3XsC8+dvAGDr1lzy8ws1wquIBEZ3ZieJLVt2M3Dg+4waNQd3OPro2gwb1p2ePY8LujQRqeAUFElg8+ZdHH/8CNav3056eiVuueUM7r77HKpXzwi6NBERBUUyOOSQqnTrdizLl//IyJE9OOkkDeAnIslDQRGA3Nx8Bg2aQfv2R9O+fWMAhg3rTpUq6RrAT0SSjoIiwT74YDV9+kxk+fIfadkyi4UL+5CWVolq1SoHXZqISLEUFAny/fc7uOWWd3nppQUAtGiRxYgRPUhL09VMIpLcFBRxtncAvwEDprBly26qVEnnrrvacdttZ5GRofGZRCT5pV5QbPky6ApKZOvW3dx55wds2bKbLl2OYfjw7hxzTN2gyxIRiVnqBUXuT6HfTboHW0cUO3bsIT29EpmZ6RxySFVGjepBQYHz298erwH8RCTlpOYJ8ibd4dcTg66iWOPHL+P440fw2GMz9j138cXHc8klJygkRCQlpWZQJGFIrFmzlQsvHMsFF4xlzZqtTJ68ksLCYkdVFxFJKakZFEkkL6+AJ574mJYth/PWW8uoWTODoUO7MnXq1bonQkTKhdTro0giGzfupGPHv7FgQWgAv9/+9niGDOlCw4a1Aq5MRKTsKCgOQr16VcnKqkaTJnUYNqw73bs3C7okEZEyp6AoAXfn5ZcXctppDTnuuHqYGS+9dBG1a1fRndUiUm6pjyJGy5ZtpFOnv9Or17/o23ci7qGO6iOOqKmQEJFyTS2KA9i9O59HHpnOo4/OYM+eAurVq8rvftcq6LJERBJGQRHFlCmr6NNnIitWbALgD3/4BY891pl69aoFXJmISOIoKPZjw4bt9Oz5Crm5BRx/fH1GjepBu3ZHB12WiEjCKSgiFBY6ZmBmHHZYDR544FwKC52bbz5DA/iJSIWlzuywefPWc+aZz+0bBhzg9tvP4o47zlZIiEiFVuFbFNu25XLvvR8xdOhMCgud3NwCfve7VhqXSUQkrMIGhbvz5ptLuf76d1i79icqVTJuuKEtDzxwrkJCRCRChQyKjRt3cs01bzFhwnIAsrMb8OyzPTnllCMCrkxEJPlUyKCoWTODFSs2UatWJg8/fB45OdmaklREZD8qTFDMmLGGFi2yqFevGpmZ6YwdezGHHlqdI46oGXRpIiJJrdx/jf7xx5386U/jOfvsvzJgwJR9z7dufbhCQkQkBuW2ReHu/O1v87n11vfYuHEnlStXokGDmri7OqtFREqgXAbF0qUbycmZwNSpXwPQoUNjRo7sQYsWWQFXJiKSespdUKxd+xOtW49iz54CsrKq8eSTv6RXL90XISJSWuUuKI48sha9erWiUiXj0Uc7Ubdu1aBLEhFJaSkfFOvWbeOmmyaTk5NNhw6NARg9+nzNVy0iUkZSNigKCgoZOXIOd975AT/9lMuKFZuYPftPmJlCQkSkDMX18lgz62pmy8xshZndUczrZmZPh19fYGanxLLfzz5bx+mnP0f//m/z00+5nH/+cbzxxiXqhxARiYO4tSjMLA0YDnQG1gKzzWy8uy+OWK0b0Cz80xYYGf69X99sqUWbNn+hsNA58shaPPNMNy64oLlCQkQkTuLZojgNWOHuq9x9DzAWuKDIOhcAf/OQT4E6ZhZ1wKVNO6tiBjfffDpLllzHhRe2UEiIiMRRPPsoGgLfRCyv5X9bC8Wt0xBYF7mSmV0LXBtezIV7Fw0eDIMHl23BKSgL2Bh0EUlCx+JnOhY/07H4WfPSbhjPoCjua76XYh3cfTQwGsDM5rh79sGXl/p0LH6mY/EzHYuf6Vj8zMzmlHbbeJ56WgscFbF8JPBdKdYREZEAxTMoZgPNzKyJmWUAlwHji6wzHrgqfPXT6cBWd19XdEciIhKcuJ16cvd8M+sHTAbSgOfd/Qszywm/PgqYBHQHVgA7gWti2PXoOJWcinQsfqZj8TMdi5/pWPys1MfC3P+nS0BERGSfcj8fhYiIHBwFhYiIRJW0QRGv4T9SUQzH4srwMVhgZh+bWesg6kyEAx2LiPXamFmBmf0mkfUlUizHwsw6mNk8M/vCzKYmusZEieH/kdpm9m8zmx8+FrH0h6YcM3vezL43s0X7eb10n5vunnQ/hDq/VwJNgQxgPnB8kXW6A28TuhfjdGBm0HUHeCzOBA4JP+5WkY9FxHofELpY4jdB1x3gv4s6wGKgUXj50KDrDvBYDAQGhR/XBzYBGUHXHodjcQ5wCrBoP6+X6nMzWVsUcRn+I0Ud8Fi4+8fuvjm8+Cmh+1HKo1j+XQD0B94Avk9kcQkWy7G4Ahjn7msA3L28Ho9YjoUDNS003k8NQkGRn9gy48/dpxH62/anVJ+byRoU+xvao6TrlAcl/Tv/j9A3hvLogMfCzBoCFwGjElhXEGL5d3EccIiZfWRmc83sqoRVl1ixHIthQEtCN/QuBG5w98LElJdUSvW5mazzUZTZ8B/lQMx/p5mdSygozo5rRcGJ5Vg8BQxw94JyPlhkLMciHTgV6AhUBT4xs0/dfXm8i0uwWI5FF2AecB5wDPCemU1395/iXFuyKdXnZrIGhYb/+FlMf6eZtQLGAN3c/ccE1ZZosRyLbGBsOCSygO5mlu/ubyakwsSJ9f+Rje6+A9hhZtOA1kB5C4pYjsU1wKMeOlG/wsxWAy2AWYkpMWmU6nMzWU89afiPnx3wWJhZI2Ac0KscfluMdMBj4e5N3L2xuzcGXgf6lsOQgNj+H3kLaGdm6WZWjdDozUsSXGcixHIs1hBqWWFmhxEaSXVVQqtMDqX63EzKFoXHb/iPlBPjsbgHqAeMCH+TzvdyOGJmjMeiQojlWLj7EjN7B1gAFAJj3L3YyyZTWYz/Lh4EXjCzhYROvwxw93I3/LiZvQp0ALLMbC1wL1AZDu5zU0N4iIhIVMl66klERJKEgkJERKJSUIiISFQKChERiUpBISIiUSkoJCmFR36dF/HTOMq628vg/V4ws9Xh9/rMzM4oxT7GmNnx4ccDi7z28cHWGN7P3uOyKDwaap0DrP8LM+teFu8tFZcuj5WkZGbb3b1GWa8bZR8vABPc/XUz+yXwhLu3Ooj9HXRNB9qvmb0ILHf3h6KsfzWQ7e79yroWqTjUopCUYGY1zOz98Lf9hWb2P6PGmtkRZjYt4ht3u/DzvzSzT8LbvmZmB/oAnwYcG9725vC+FpnZjeHnqpvZxPDcBovM7NLw8x+ZWbaZPQpUDdfxcvi17eHf/4j8hh9uyVxsZmlm9riZzbbQPAG9YzgsnxAe0M3MTrPQXCSfh383D9+l/ABwabiWS8O1Px9+n8+LO44i/yPo8dP1o5/ifoACQoO4zQP+RWgUgVrh17II3Vm6t0W8Pfz7FuDO8OM0oGZ43WlA9fDzA4B7inm/FwjPXQH8FphJaEC9hUB1QkNTfwGcDFwM/CVi29rh3x8R+va+r6aIdfbWeBHwYvhxBqGRPKsC1wJ3hZ/PBOYATYqpc3vE3/ca0DW8XAtIDz/uBLwRfnw1MCxi+4eB34Uf1yE07lP1oP976ye5f5JyCA8RYJe7/2LvgplVBh42s3MIDUfREDgMWB+xzWzg+fC6b7r7PDNrDxwPzAgPb5JB6Jt4cR43s7uAHwiNwtsR+JeHBtXDzMYB7YB3gCfMbBCh01XTS/B3vQ08bWaZQFdgmrvvCp/uamU/z8hXG2gGrC6yfVUzmwc0BuYC70Ws/6KZNSM0Gmjl/bz/L4Ffmdmt4eUqQCPK5xhQUkYUFJIqriQ0M9mp7p5nZl8R+pDbx92nhYOkB/B3M3sc2Ay85+6Xx/Aet7n763sXzKxTcSu5+3IzO5XQmDmPmNm77v5ALH+Eu+82s48IDXt9KfDq3rcD+rv75APsYpe7/8LMagMTgOuApwmNZfShu18U7vj/aD/bG3Cxuy+LpV4RUB+FpI7awPfhkDgXOLroCmZ2dHidvwDPEZoS8lPgLDPb2+dQzcyOi/E9pwEXhrepTui00XQzawDsdPeXgCfC71NUXrhlU5yxhAZja0doIDvCv/vs3cbMjgu/Z7HcfStwPXBreJvawLfhl6+OWHUboVNwe00G+lu4eWVmJ+/vPUT2UlBIqngZyDazOYRaF0uLWacDMM/MPifUjzDU3X8g9MH5qpktIBQcLWJ5Q3f/jFDfxSxCfRZj3P1z4CRgVvgU0J3An4vZfDSwYG9ndhHvEprbeIqHpu6E0Fwii4HPzGwR8CwHaPGHa5lPaFjtxwi1bmYQ6r/Y60Pg+L2d2YRaHpXDtS0KL4tEpctjRUQkKrUoREQkKgWFiIhEpaAQEZGoFBQiIhKVgkJERKJSUIiISFQKChERier/A8f/v9BqxAgVAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# call the function with the best parameters\n",
    "\n",
    "auc_score, prc_socre = create_model(train[\"text\"], train[\"label\"], test[\"text\"], test[\"label\"], \\\n",
    "            tokenizer=None, model_type='svm',  min_df = 1, stop_words=None,\\\n",
    "                                    print_result=True, algorithm_para=0.5)\n",
    "# Please answer the questions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 3. How many samples are enough? \n",
    "\n",
    "This task is to help you understand the impact of sample size on classifier performance. \n",
    "\n",
    "- Define a function `sample_size_impact(train_docs, train_y, test_docs, test_y)` with the input as defined above. \n",
    "    \n",
    "    \n",
    "- This function does the following:\n",
    "    - Starting with 500 samples, in each round you build a classifier with 500 more samples. i.e. in round 1, you use samples from 0:500, and in round 2, you use samples from 0:1000, …, until you use all samples. \n",
    "    - In each round, call the `create_model` function with `model_type='svm'` and  `model_type='nb'` and `stop_words='english'`. Keep other parameters as default. Record the AUC score of each model type and the sample size.\n",
    "    - For each model type, plot a line chart show the relationship between sample size and the AUC score. \n",
    "    - This function has no return.\n",
    "    \n",
    "    \n",
    "- Write your analysis on the following:\n",
    "    - How sample size affects each classifier’s performance? \n",
    "    - If it is expensive to collect and label samples, can you decide an optimal sample size with model performance and the cost of samples both considered? \n",
    "    - How is performance of SVM classifier compared with Naïve Bayes classifier?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Show the impact of sample size\n",
    "\n",
    "def sample_size_impact(train_docs, train_y, test_docs, test_y):\n",
    "    \n",
    "    svm_s=[];\n",
    "    nb_s=[];\n",
    "    col=[];\n",
    "    \n",
    "    for i in range(500,len(train_docs)+1,500):\n",
    "        col.append(i);\n",
    "        \n",
    "        auc_score,pcr_score= create_model(train_docs[0:i], train_y[0:i], test_docs, test_y,tokenizer=None, model_type='svm', stop_words = None, min_df = 2, print_result=False, algorithm_para=1.0)\n",
    "        \n",
    "        svm_s.append(auc_score);\n",
    "        \n",
    "        auc_score,pcr_score= create_model(train_docs[0:i], train_y[0:i], test_docs, test_y,tokenizer=None, model_type='nb', stop_words = None, min_df = 2, print_result=False, algorithm_para=1.0)\n",
    "        \n",
    "        nb_s.append(auc_score)\n",
    "        \n",
    "    y_labels=['', 0.84, 0.86, 0.88, 0.90, ''];\n",
    "\n",
    "    plt.figure();\n",
    "    plt.plot(col,svm_s, color='darkorange', lw=2);\n",
    "    plt.plot(col,nb_s, color='blue', lw=2);\n",
    "    plt.xlim([0, 12000]);\n",
    "    plt.xlabel('sample size');\n",
    "    plt.ylabel('AUC');\n",
    "    plt.show(); \n",
    "        \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Test the function\n",
    "\n",
    "# now we use all the data\n",
    "train = pd.read_csv(\"train5.csv\")\n",
    "test = pd.read_csv(\"test5.csv\")\n",
    "\n",
    "sample_size_impact(train[\"text\"], train[\"label\"], test[\"text\"], test[\"label\"])\n",
    "\n",
    "\n",
    "# Briefly answer the questions\n",
    "#1) When sample size increases->Increase in AUC for SVM as compared to NB\n",
    "#2) Optimal sample size is hard to determine along with costs\n",
    "#3)The AUC for NB first rises until the sample size reaches about 2000, but after that it tends to remain fairly stable or fluctuate just slightly. When SVM is examined, the AUC rises along with the spike in sample size, making it superior to NB."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 4 (Bonus): How to improve sentiment classification?\n",
    "\n",
    "- Can you tune other parameters of TfidfVectorizer or SVM model to further improve the classification performance? Try to beat the model you get in Task 2!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    " # add your code here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}