int20_tf-idf.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import re"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk\n",
    "#nltk.download('stopwords')\n",
    "from nltk.corpus import stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import naive_bayes, svm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = pd.read_csv(\"train_int.csv\")\n",
    "train_data.drop(columns=['id'], inplace=True)\n",
    "test_data = pd.read_csv(\"test_int.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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": [
    "cl = ['Positive', 'Negative']\n",
    "positive = train_data[train_data.sentiment == 1].sentiment.sum()\n",
    "negative = train_data[train_data.sentiment == 0].sentiment.count()\n",
    "counts = [positive, negative]\n",
    "barlist = plt.bar(cl, counts)\n",
    "barlist[0].set_color('#800000')\n",
    "barlist[1].set_color('g')\n",
    "plt.title('Classes distribution in the train data', fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "REPLACE_BY_SPACE_RE = re.compile('[/(){}\\[\\]\\|@,;#]')\n",
    "BAD_SYMBOLS_RE = re.compile('[^0-9a-z +_]')\n",
    "STOPWORDS = set(stopwords.words('english'))\n",
    "\n",
    "def text_prepare(text):\n",
    "    \"\"\"\n",
    "        text: a string\n",
    "        returns: modified initial string\n",
    "    \"\"\"\n",
    "    text = text.lower()\n",
    "    text = REPLACE_BY_SPACE_RE.sub(\" \", text)\n",
    "    text = BAD_SYMBOLS_RE.sub(\"\", text)\n",
    "    text = text.split()\n",
    "    text = \" \".join([word for word in text if not word in STOPWORDS])\n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data['review'] = train_data['review'].apply(text_prepare)\n",
    "test_data['review'] = test_data['review'].apply(text_prepare)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data['review_lenght'] = np.array(list(map(len, train_data.review)))\n",
    "median = train_data['review_lenght'].median()\n",
    "mean = train_data['review_lenght'].mean()\n",
    "mode = train_data['review_lenght'].mode()[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtAAAALZCAYAAACeWBZUAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAABXIElEQVR4nO3de0CUVcLH8d8MA5oo3gIyNEsTzCzxrl0wLdEN0Uwry7TLvt3WS9muq5lpWamZaV7SLddqM9tyvVC6ppVdNM17apahm5qlBl5BQGFgnvcPY3JkwDnKMFy+n7+YM8/lPIcH/c2Z85xjsyzLEgAAAACf2ANdAQAAAKAsIUADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQQAUwfPhwxcTE6IsvvvD6fq9evRQTE6O+fft6fT8pKUkxMTGaNGmSP6vp1qpVK3Xq1KlEzlWe5f/ed+zYEeiqFOnFF19UTEyM1q1b5y7r1KmTWrVqdV7HS01N1YIFC3zadt26dYqJidGLL75YLOc+lz179ujjjz/2KIuJiVGPHj38cj4A/kGABiqAtm3bSpK2bNlS4L3jx4/rhx9+kN1u19atW5WZmVlgm02bNkmS2rdv79d6onjdcsstGjhwoC6++OJAV8VY//799fDDDxvvd+TIEXXt2lUrVqzwafuoqCgNHDhQN954o/G5TP34449KTEzU5s2bPcoHDhyoPn36+P38AIqPI9AVAOB/+QF669atBd775ptv5HK51KVLFy1fvlzr169Xx44dPbbZtGmTKlWqpBYtWpRIfVE8brnlFt1yyy2BrsZ5uf/++89rv5MnT3r9EFiYunXratCgQed1LlNpaWlyOp0Fykvq/ACKDz3QQAVw6aWXql69etq2bZtcLpfHe2vWrJHD4dCAAQMkSatXr/Z4//jx49q9e7eaN2+uSpUqlVidAQAorQjQQAXRtm1bZWZmateuXR7lq1ev1rXXXquYmBjVq1dPa9as8Xh/8+bNsiyrwPCNpUuXqk+fPoqNjVXz5s3Vp08f/fe///XY5tdff1VMTIymTJmiF154QbGxsWrbtq17DOjRo0f13HPP6cYbb1SzZs10//33Kzk52Wv9v/76a913331q3769rr32WiUmJur1119XTk7OOa+9X79+iouL0/79+/Xoo4+qefPmuu666zR06FAdOHCgwPYZGRmaOHGibrnlFjVt2lQ33nijRo8erSNHjnhslz/GeNu2bbr11lt1zTXXqE+fPrIsy2s9pk2bppiYGH3zzTe644471LRpU3Xp0sXdY3ro0CE9++yziouLU9OmTdWpUye9/PLLysjIcB/jhRdeUExMjFatWlXg+Fu2bFFMTIyeffZZj/qdPQb6m2++0QMPPKCWLVsqNjZWd911l5YtW+Z+f9euXYqJidGwYcM89ktOTlZMTIxuuukmj3KXy6W2bdsWOob+TPPnz1f37t3VrFkzxcfH6/333/e6nbdxyEuWLFGfPn3UunVrNW/eXL169dJ7773nbu+FCxfq5ptvliStWLFCMTExWrhwoaTT44yHDx+uf/zjH2rVqpVatWqlt99+2+sY6Hw7d+7Ufffdp2bNmun666/XqFGjCtwD/fr1U0xMjNLT0z3K8+/9v/zlL5JO/+779+8vSXrnnXc8xnx7GwN94sQJTZgwwX0PXnfddfrrX/+qPXv2eGy3cOFC9z01e/ZsxcfHq2nTprrllls0c+ZM5eXlFfKbAHAhCNBABdGmTRtJnuOgf/75Z+3fv1/XXXedJOm6667TTz/9pJSUFPc23sY/v/TSSxoyZIh+/fVXdevWTQkJCfr111/15JNP6uWXXy5w7nnz5unjjz/W3XffrdjYWMXGxiozM1P33nuv3nvvPTVs2FB33XWXjh8/rn79+unUqVMe+2/cuFGPPvqodu/erVtvvVX33nuvgoKCNGnSJHdYPJdTp06pf//+2rNnj/r06aOmTZvqo48+Up8+fTyu98SJE7r77rs1a9Ys1a1bV/3791fz5s01b9483XHHHUpNTS1w7Mcee0yXXXaZ+vTpo7Zt28pmsxVZl7/97W+qXLmy+vXrp7Zt2yo0NFQHDhxQ79699f777+vqq6/W/fffryuuuEL//Oc/1a9fP2VlZUmSunfvLkkFHkST5P4Ak7+NN//5z3/0wAMPKDk5WbfeeqvuuusuHTlyRI8//rj+8Y9/SJIaNWqkSy+9VGvXrvXYN//1wYMH9csvv7jLt23bpuPHjxcI1md79dVX9fTTTysjI0O9e/dW48aNNWbMGK/X4u3a/vrXv+rYsWPq2bOn7rrrLqWnp+u5557TjBkzJElXXXWVO6ReccUVGjhwoK666ir3MVatWqVZs2bptttu0w033KBmzZoVer5Tp07p3nvv1YkTJ9S3b181aNBAH3zwge655x6PDzS+atOmjXr27ClJatasmQYOHKioqCiv2x47dkx33HGHZs+erdq1a6tv376KjY3V0qVL1bt3b69DsV5++WVNnz5dLVu2VN++fXXq1Cm9+uqrmjp1qnFdAfjAAlAh/Pbbb1Z0dLQ1fPhwd9ncuXOt6Ohoa8OGDZZlWdZ///tfKzo62lqwYIF7m7vuustq2bKllZuba1mWZW3YsMGKjo62brvtNuvIkSPu7Y4cOWJ169bNio6OttavX29ZlmX98ssvVnR0tBUTE2Pt2LHDoz5TpkyxoqOjrWnTprnLnE6n9eSTT1rR0dFWx44d3eWDBg2yoqOjrX379rnLcnJyrB49elhXXXWVdeLEiSKv/d5777Wio6OtXr16WSdPnnSXz54924qOjraGDRvmLnv22Wet6Oho69133/U4xmeffWZFR0dbgwcPdpcNGzbMio6OtgYOHFjk+fNNnTrVio6Otm6//XYrLy/P472HHnrIiomJsb744guP8n/9619WdHS09dJLL7nL4uPjrdatW1s5OTnusry8POuGG26wOnXqVKB+P/zwg2VZlnXw4EGradOm1p/+9Cfr6NGj7u1Onjxp3XXXXVbjxo2t5ORky7Is65lnnrGio6Ot3bt3u7d79NFHrdjYWCs6OtpauHBhgevauXNnode+Z88eq0mTJlaPHj2stLQ0d/nnn39uxcTEWNHR0dbatWvd5R07drRatmzpft2zZ08rNjbW43d94sQJ6/rrr7fatWtnuVwuy7L+uOcee+wxj/NHR0db0dHR1ooVKzzK165da0VHR1svvPCCx7mjo6OtQYMGefyeJk6caEVHR1uTJk1yl+XfW2deU2H18Hau/Lp1797d/fqpp56yoqOjrcmTJ3ts9+WXX1oxMTFWfHy8++9xwYIFVnR0tNWyZUtr7969Hue/+uqrreuuu84CUPzogQYqiMjISF1++eUePdBr1qxRlSpVFBsbK0lq166dbDabexhHTk6Ovv/+e7Vu3VpBQUGS5P5K/O9//7tq1arlPlatWrX017/+VZIKTCFWv359NW7c2KPsv//9r8LCwvTII4+4yxwOh4YNG1agBzd/3PZ3333nLgsODtasWbO0bt06Va1a1ac2ePLJJ1W5cmX36/vuu09RUVFavny5cnJylJubq6SkJDVq1KjAcISbb75ZLVq00KefflqgBzI+Pt6n8+fr3Lmz7PY//vlNTU3VypUr1aFDhwK9uPfee6/q1KmjRYsWucsSExOVlpbmMV5948aNSk1NVbdu3Qo970cffaScnBwNHjxYNWvWdJdXrlxZgwcPlsvlcp+nQ4cOkk4P95CkvLw8bdy4UT179lRwcLA2bNjg3v/rr79WVFSUGjVqVOi5ly1bptzcXD366KMKCwtzl3fs2FE33HBDofvlsyxLp06d8hiCVLVqVc2fP18rVqw4Z69//nXmX9e52Gw2DR061OP3NGjQIFWtWlWLFy/26RjnIycnR//9738VFRWlwYMHe7zXoUMHxcfHa+/evdq4caPHe/Hx8apfv777dd26ddWwYUMdPnxY2dnZfqsvUFExCwdQgbRt21bz5s1Tenq6QkNDtW7dOrVp00YOx+l/CmrVqqXGjRtr/fr1kk5/NZ+Tk+MxfOPHH3+U3W5Xy5YtCxw/v+zHH3/0KK9bt67H61OnTmnv3r1q06aNgoODPd6LiIhQ3bp1PR52vOOOO/TZZ59pyJAhmjJlim688UbFxcWpXbt2CgkJ8enabTZbgTG1QUFBuvrqq/XJJ59o3759sixLWVlZysvL07Rp0wocIzs7W3l5eUpOTva4/rOv71zO3v6HH36QZVk6fvy41/MGBwfr4MGDSklJUWRkpHr06KFp06Zp6dKl7sDty/CN7du3Szodis8eC58/RCT/d9e+fXuFhIRo7dq1uueee/T9998rPT1d119/vb777jt3gDt+/Li+++67c07Dln/cpk2bFnivefPmXsd0n+muu+7S6NGj1adPH8XExCguLk4dOnRQy5YtPUJuUS655BL3B8FziYiIUL169TzKQkJC1LhxY23cuFEnTpxQtWrVfDqWiT179ujUqVNq0aKF1+tq2bKlli9frh9//NE9u44kXX755QW2za9fTk4ODwADxYwADVQgbdu21QcffKAtW7YoLCxM6enpBR4ObN++vd58803t27fPPV/tmdtkZGSoUqVKXoNrtWrVdNFFF+nkyZMe5Wf/552WliZJCg0N9VrP6tWr69ixY+7XHTp00DvvvKPZs2drzZo1mjNnjubMmaMaNWpo4MCB6tev3zmvvWbNml7rnD9H8okTJ9yhfffu3Zo+fXqhx8qvf74ze7V9cfb2+Q+gbdmyxetc3fmOHz+uyMhI1atXT82bN9eKFSuUk5Mju92uTz75RFdffbUaNmxY6P4nTpyQpEIf3JP+uLYqVaqodevWWrdunVwul9auXSu73a7WrVtr06ZNmj17tlJTU7Vx40bl5eWds2c3/xq9/c5r1KhR5L6S1KdPH9WuXVvvvPOONm3apOTkZM2aNUuRkZEaPny4br311nMew+T3VNjc2fn1z8rK8kuAzv92o7BjR0RESFKB5wS83dv5vfJWIQ+1Ajh/BGigAsl/kHD79u3u3q3CAvTmzZu1efNmhYeHe3w1HxoaqpMnTyo9Pd3jq3jpdA/tqVOnPIYHeFO9enVJfwS6s+X3hp5d9zZt2igrK0sbN27Ul19+qUWLFumFF17QZZddds4AV9jX2PnBrmbNmu5Q0qNHD02YMKHI4xWnKlWqSJL+8pe/6PHHH/dpn8TERI0ZM0YrV65U5cqVdfTo0XMuPJJ/ns8++6xA76o3cXFxWr16tXbs2KH169ercePGCgsLU9u2bTV79mxt3LhRq1at0kUXXaR27doVeaz8eyUjI6PA/eHrvM2dO3dW586dlZ6ernXr1unzzz/X4sWL9de//lVXXnmloqOjfTqOL86eVSNfamqqbDZbgXv/7Okhzw64vsoP6Gc+2OqtXr586ADgP4yBBiqQ8PBwNWjQQN9//702bdqkiy++WDExMR7btG7dWsHBwUpOTtbWrVs9viaW5B7LnD87x5k2bdoky7J05ZVXFlmPypUrq2HDhvrhhx8KBI309HTt27fPo+xf//qXXn31VUmnQ2BcXJxGjRql0aNHF1qXs2VmZuqnn34qUL5161bVrFlT9erV0xVXXKGQkBB9//33Xnvt3n77bc2YMcOjd7w45P8O8odYnG3q1Kl64403PKbsu/XWWxUcHKzPP/9cy5Ytk91uV0JCgk/nOXMseb69e/fqpZde0ueff+4ui4uLk3R6rPyWLVvUunVrSaeHETgcDq1fv15ff/212rVrd84hAldffbUk77+rwq47X05OjmbOnKm3335b0ukw3rlzZ40bN06PPfaYXC6Xvv32W0nyaSy0Lw4cOFBgyrrjx49r7969atCggS666CJJf/T8nv2ty9n3sK91a9CggSpVqqTvvvvO6xSN+WPPz/U3BsC/CNBABdO2bVt999132rJli9dew4suukixsbH64osvdPTo0QI91LfffrskadKkSTp69Ki7/OjRo+5e27PntPWmZ8+eysrK0sSJE91h1bIsvfLKK8rNzfXY9uuvv9Y//vGPAsMb9u/fL+n0QjG+mDhxokcoefPNN/XLL7+oZ8+eCgoKUqVKlXTrrbfqf//7n9566y2PfdetW6cJEyZowYIF7h704lKvXj21bt1aK1eu9JiPWZKSkpL02muvadWqVR5f09esWVM33HCDVq5cqS+++ELt2rVzf71fmO7duysoKEivvvqqDh065C7Pzc3V888/rzfffFPHjx93lzdo0ECXXXaZ/v3vf+vEiRPubzCqVq2qJk2aaMmSJUpNTT3n9HXS6cBfqVIlzZw50+PcGzdu9Ajt3oSEhGjJkiWaMmWKx/R5UsF7IH88v7cV/0zk5eXptddec7+2LEuTJk3SyZMn1atXL3f5FVdcIUn64osv3GXZ2dmaPXt2gWP6UreQkBAlJCQoNTW1wBR0K1eu1Mcff6z69euzKigQYAzhACqYtm3b6t///rckued/Plv79u3d/3mfHaBbt26tBx54QG+99Za6d+/uXvb7iy++0KFDh/TQQw+5eyqLct999+nzzz/XnDlztH37djVr1kxbtmzRzp07Vbt2bY9tBw0apHXr1ql///7q2rWrIiMj9b///U9ffPGFGjZsWOSDc2fasGGDevXqpfbt2+unn37S119/rUaNGrlXYZSkYcOG6dtvv9VLL72kFStW6Nprr1VKSoo++eQTORwOjR071ueH1kyMGTNGffv21eOPP664uDg1atRIe/bs0ZdffqkaNWq4e9vP1L17d3dwy58BpSiXX365hg4dqvHjx6tbt27q1KmTqlevrpUrV+qnn35Sx44dC7RlXFyc3n33Xff453xt2rTRtm3bJMmnmS2ioqI0bNgwjRkzRj179tQtt9yijIwMLVu2THXq1PHaY3umJ598UgMGDFDPnj3VtWtXVa9eXdu3b9fatWvVpk0bXX/99ZL+GOu+bt06jRs3Tp07dy7w8KgvateuraSkJO3atUtNmzbV5s2btWXLFrVq1co917Qk9e7dW++9957Gjh3r/jZjxYoVqlatmnvITL7IyEhJp+fwrlKlinr27Ol15pKhQ4dq8+bNmjVrljZs2KDmzZvrl19+0eeff67Q0FC9/PLLxdbTDuD80AMNVDBnLvRRVICWTveMelvsYfjw4Xr55ZcVFRWlxYsX6+OPP9YVV1yhadOm6W9/+5tP9QgJCdGbb76phx9+WCkpKe5Q/+abbyo8PNxj22uvvVbvvvuurr/+eq1du1ZvvfWWkpOT1b9/f82dO7dAUCnMP//5T0VEROiDDz7Qrl271L9/f7333nse0+DVqlVL8+bN04MPPqiUlBTNmTNHGzduVKdOnTRv3rwCQ1qKS4MGDbRw4ULdeeedSk5O1jvvvKPk5GT16NFD8+fP9/qV/c0336yqVauqUqVKPk+l98ADD+iNN95Q48aN9cknn+iDDz6Qw+HQ8OHDNXXqVHcvab78YRwxMTEePe/57RATE6M6der4dO6+ffvqtddec0/Lt3HjRg0ePNinFQxvvvlmzZ49W02bNtUXX3yhd955R7/99psGDBigN954w/2hJiQkRKNGjVL16tX13nvvFVgMxlfh4eF666235HQ69e677+rgwYN66KGH9M9//tNj5pjGjRvrjTfeUNOmTfXxxx/ro48+Uvv27fX2228XmPEjKipKTzzxhGw2m+bOnev+AHK2M+/BQ4cO6d1339V3332n2267TQsXLixyARgAJcNm8XgugHKuX79+Wr9+vTZs2FDg4S8AAEzRAw0AAAAYIEADAAAABgjQAAAAgAHGQAMAAAAG6IEGAAAADJT7eaCPHMmQy1WxOtl/mTBOklTv708VuV3NmlV07FjBJZNNVK9+qyQpLW3pBR3nfN2WdPr8SbcV//mLo33KM9qncLRN0WifotE+RaN9CkfbFC08vFqxHYse6ArM4Qg690YVGO1TNNqncLRN0WifotE+RaN9CkfblBwCNAAAAGCAAA0AAAAYIEADAAAABsr9Q4QAAABlXV5ero4dO6Tc3JxCt0lNtcvlcpVgrUonhyNENWuGKyjIfzGXAA0AAFDKHTt2SJUrV1Fo6CWy2Wxet3E47MrNrdgB2rIsZWam69ixQ7r44jp+Ow9DOAAAAEq53NwchYaGFRqecZrNZlNoaFiRPfXFgQANAABQBhCefVMS7USABgAAAAwQoAEAAAADBGgAAADAALNwAAAAwMjmzRv1zjtvyrKkAwd+1U033azQ0FCtWvWVLMvSxIlTtHNnsmbP/odyc3NVp06Uhg17WtWr19Dnn3+m999/V9nZ2crOztbw4SMVG9tCAwc+rCZNrtbWrVt0/PgxPfHEULVvf32gL9UrAjQAAEAZ8sGP7+nfP75boNxmkyzrwo59d+N7dVfje3za9ocfvtecOR+oevUaSkzsrAEDntDs2XM0duxzSkpaoJUrv9TUqf9QWFiYkpIWaObMafr735/Whx8u0IQJr6pGjRpasuRDvffeHMXGtpAkOZ25ev31t/T11ys1a9ZMAjQAAADKjwYNGioy8hJJUvXqNdSqVRtJUmTkJVq9epVSUn7T4MGPSpJcrjyFhVWX3W7X2LEva/XqVdq372d9++0m2e1/jChu27a9+9gnTqSX8BX5jgANAABQhtzV+B6vvcQlvZCKw+EZI4OCgtw/u1x5uvbaZnrppcmSpOzsbGVlZSkrK0v/93/91aXLrWrWrLkaNrxSCxbMc+8XEhIi6fRUdNaFdqf7EQ8RAgAAoFg1adJU33//nfbt+1mS9Pbb/9SMGVP0yy/7ZLfb1b//g2rZsrXWrl1TJpcfpwcaAAAAxapWrdoaPnyURo16Si5XnsLDIzVq1BhVrVpNV14ZrXvu6a3KlSsrNraFfvvtYKCra4wADQAAACMtWrRSixat3K/nz1/s/vnPf37E/fMNN8QV2Pe558Z6vH7iib9JkqZPf8NdVqfOpR7HLG0YwgEAAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAAJi8+aNGjjw4UBXwxgBGgAAADDASoQAAABlSPqa1Ur7emWBcpvNJsuyLujY1W+IU9h1159zu82bN+qdd96UZUkHDvyqm266WaGhoVq16itZlqWJE6dox44fNGvWTFmWS5deGqWhQ0eoVq3aWr9+raZOnaSQkBDVr3+5+5i//vqLJk4cp/T0NFWqVFlDhgxVdHTjC7oef6EHGgAAAMZ++OF7jRgxSnPmzFNS0nzVqFFTs2fP0ZVXNlJS0gK9/PJYjRs3Uf/61/u65ppmmjRpgnJycvTii6P1wgsv6c0331WlSpXcx3vxxdH6y18G68035+rvf39ao0ePCODVFY0e6PLIZpMk2e02d5HLdWGfSAEAQOkQdt31XnuJHQ67cnNdJVaPBg0aKjLyEklS9eo11KpVG0lSZOQlWr16la666mrVqXOpJKl799s1Z87b2r37f6pdO1yXX36FJOlPf+qmWbNmKisrSzt2/KCxY8e4j3/y5EmlpR1X9eo1SuyafEWALmfsdptSj2UpN8/S0qTtkqSw0BD17RxNiAYAAMXG4fCMkUFBQe6fLcszyFuWpby8PEk2j/fy93G5XAoJqaS3337P/V5qaorCwqr7oeYXjiEc5VBunqW8PJfSMrKVlpGt9MycQFcJAABUIE2aNNUPP3yngwcPSJI++mihWrRoqSuvbKRjx45p166dkqTPPlsuSapatarq1q2n5cuXSpI2bFirAQNK7+wc9EADAACgWNWsWUtDhz6tESP+JqczV5dccomGDx8lh8OhZ599US+8MEpBQUEeDwmOHv2CXn55rN577x05HMEaM2asbDZbEWcJHAI0AAAAjLRo0UotWrRyv54/f7H75z//+RH3zzfcEFdg39jYFvrXv94vUF6//uWaPv2NYq6pfzCEAwAAADBAgAYAAAAMEKABAAAAAwRoAAAAwIBfA/T06dOVkJCghIQETZgwQZL01FNPKT4+Xj169FCPHj306aefSpLWrFmjxMRExcfHa/Lkye5j7NixQ7169VKXLl309NNPKzc3159VBgAAAIrktwC9Zs0aff3111q0aJGSkpL0/fff69NPP9X27dv17rvv6sMPP9SHH36ozp0769SpUxoxYoRmzJihpUuXavv27frqq68kSUOHDtUzzzyj5cuXy7IszZs3z19VBgAAAM7JbwE6PDxcw4cPV0hIiIKDg9WwYUMdOHBABw4c0DPPPKPExERNnTpVLpdL27ZtU/369VWvXj05HA4lJiZq2bJl2r9/v06dOqXY2FhJ0u23365ly5b5q8oAAAAoQZs3b9TAgacXTBk//nn9+OMPAa6Rb/w2D3SjRo3cP+/du1dLly7Ve++9p/Xr12vMmDGqUqWKHnnkEc2fP19VqlRReHi4e/uIiAilpKQoNTXVozw8PFwpKSlG9ahdu+qFX0wZY7dJLptNwcGnf73BwUGFtkN4eLULPFtQMR3n/AQH+/f8gbqusoL2KRxtUzTap2i0T9EqYvukptrlcJy739OXbUqToCC7bDabHA67Ro4cXWzHtdvtfr1P/L6Qyq5du/TII49o2LBhatCggV577TX3e/369VNSUpK6du1aYD+bzSbLsryWmzhyJEMuV8HjlFd2u00u6/Sa807n6fHiTmeQ13YID6+mQ4dOXND5qlfPkySlpV3Ycc6X03n6/Bd6Hd4UR/uUZ7RP4WibotE+RaN9ilZR28flcik311XkNg6H/ZzbFJfNmzfqnXfelGVJBw78qptuulmhoaFateorWZaliROnaOfOZM2e/Q/l5uaqTp0oDRv2tKpXr6H169dq6tRJCgkJUf36l8uyLOXmujRw4MN68MGHde21sXrllfHavfsnHT16VJddVl9jx07Q0aNHNWLE39SgQUPt3JmsWrVq6/nnxyssrHqB+rlcrgL3SXEGar8G6E2bNmnw4MEaMWKEEhISlJycrL1796pLly6SToc8h8OhyMhIHT582L1famqqIiIiCpQfOnRIERER/qwyAABAqVap0nuqXPndAuU2m+Sl79HIqVP3Kjv7Hp+2/eGH7zVnzgeqXr2GEhM7a8CAJzR79hyNHfuckpIWaOXKLzV16j8UFhampKQFmjlzmp58cphefHG0pkz5hy6//AqNH/98geNu375NDkewXn/9LblcLg0e/Ki++Wa1YmKu0v/+t0tPPTVK0dGN9fTTQ/XJJx+rd+8+F3bR58FvAfrgwYMaMGCAJk+erPbt20s6HZjHjh2rdu3aqUqVKvrggw/Us2dPNWvWTHv27NHPP/+sunXrasmSJerVq5eioqJUqVIlbdq0SS1btlRSUpLi4gouCQkAAICS1aBBQ0VGXiJJql69hlq1aiNJioy8RKtXr1JKym8aPPhRSZLLlaewsOravft/ql07XJdffoUk6U9/6qZZs2Z6HDc2toXCwqprwYJ52rdvr3799RedPHlSklSzZi1FRzf+/fxXKj09vUSu9Wx+C9CzZ89Wdna2xo8f7y7r06ePHn74Yd19993Kzc1VfHy8unXrJkkaP368Bg0apOzsbHXo0ME9rGPixIkaOXKkMjMz1aRJE/Xv399fVQYAACj1srPv8dpLXJJDOE6fzzNGBgUFuX92ufJ07bXN9NJLp6cmzs7OVlZWllJSfpNlubzuk+/rr7/SP//5uu64o49uvbW7jh8/7h7WGxIS4rGtt+G+JcFvAXrkyJEaOXKk1/f69u1boKx9+/b66KOPCpQ3btxY8+fPL/b6AQAAwD+aNGmqr776XPv2/azLLquvt9/+pw4fPqRhw0bq2LFj2rVrpxo1itZnny0vsO/GjevVqdMtSkjorsOHD2nr1m/VunWbAFxF4fz+ECEAAAAqllq1amv48FEaNeopuVx5Cg+P1KhRY+RwOPTssy/qhRdGKSgoyD0c40yJiT313HNP64svPlNwcIiuvrqpDhw4oJYtA3AhhSBAAwAAwEiLFq3UokUr9+v58xe7f/7znx9x/3zDDQWfXYuNbaF//ev9AuXTp7/h/vmddz7wet7CzlPSytZkgQAAAECAEaABAAAAAwRoAAAAwAABGgAAoAwI1JRtZU1JtBMBGgAAoJRzOEKUmZlOiD4Hy7KUmZkuhyPk3BtfAGbhAAAAKOVq1gzXsWOHlJFxvNBt7Ha7XK6SW0iltHI4QlSzZrh/z+HXowMAAOCCBQU5dPHFdYrcJjy8mg4dOlFCNarYGMIBAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGHAEugK4MHa7rcjXAAAAKF4E6DLMbrdp7qc7lZ6Z4y67NDxUEQGsEwAAQHlHgC7j0jNzlJaR7X4dFhpCgAYAAPAjxkADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGDAEegKoGTY7bZAVwEAAKBcIEBXAGGhIZqzPFnpmTkeZYP7tAhgrQAAAMomAnQFkZ6Zo7SM7EBXAwAAoMxjDDQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAb8GqCnT5+uhIQEJSQkaMKECZKkNWvWKDExUfHx8Zo8ebJ72x07dqhXr17q0qWLnn76aeXm5kqSDhw4oL59+6pr16567LHHlJmZ6c8qAwAAAEXyW4Bes2aNvv76ay1atEhJSUn6/vvvtWTJEo0YMUIzZszQ0qVLtX37dn311VeSpKFDh+qZZ57R8uXLZVmW5s2bJ0l67rnndM8992jZsmVq2rSpZsyY4a8qAwAAAOfktwAdHh6u4cOHKyQkRMHBwWrYsKH27t2r+vXrq169enI4HEpMTNSyZcu0f/9+nTp1SrGxsZKk22+/XcuWLZPT6dSGDRvUpUsXj3IAAAAgUPwWoBs1auQOxHv37tXSpUtls9kUHh7u3iYiIkIpKSlKTU31KA8PD1dKSoqOHTumqlWryuFweJQDAAAAgeLw9wl27dqlRx55RMOGDZPD4dCePXs83rfZbLIsq8B+RZWbqF27qlmFy5jg4CAFB//xa3Q4gmS3SS6bzV3ucAQV2C44OEiSFB5e7QJrUFzHOT/Fdx3eBeq6ygrap3C0TdFon6LRPkWjfQpH25QMvwboTZs2afDgwRoxYoQSEhK0fv16HT582P1+amqqIiIiFBkZ6VF+6NAhRUREqFatWsrIyFBeXp6CgoLc5SaOHMmQy1UwiJcHdrtNTmeenM5cd1lubp5clmRZlrs8NzevwHZO5+ngeejQiQuqQ/XqeZKktLQLO875cjpPn/9Cr8Ob8PBqfjlueUH7FI62KRrtUzTap2i0T+Fom6IV54cLvw3hOHjwoAYMGKCJEycqISFBktSsWTPt2bNHP//8s/Ly8rRkyRLFxcUpKipKlSpV0qZNmyRJSUlJiouLU3BwsFq1aqWlS5d6lAMAAACB4rce6NmzZys7O1vjx493l/Xp00fjx4/XoEGDlJ2drQ4dOqhr166SpIkTJ2rkyJHKzMxUkyZN1L9/f0nS6NGjNXz4cM2cOVN16tTRpEmT/FVlAAAA4Jz8FqBHjhypkSNHen3vo48+KlDWuHFjzZ8/v0B5VFSU5syZU+z1AwAAAM4HKxECAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABhyBrgACy263ebx2uawA1QQAAKBsIEBXUGGhIZoxf6uOpp30KOvbOZoQDQAAUAQCdAV2IitHaRnZga4GAABAmcIYaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAw4Ah0BVC62O22AmUulxWAmgAAAJROBGi4hYWGaM7yZKVn5niU9e0cTYgGAAD4HQEaHtIzc5SWkR3oagAAAJRajIEGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMCAI9AVgO/sdluRrwEAAOB/BOgywm63ae6nO5WemeMuuzQ8NIA1AgAAqJgI0GVIemaO0jKy3a/DQkMCWBsAAICKiTHQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABvweoDMyMtStWzf9+uuvkqSnnnpK8fHx6tGjh3r06KFPP/1UkrRmzRolJiYqPj5ekydPdu+/Y8cO9erVS126dNHTTz+t3Nxcf1cZAAAAKJRfA/TWrVt19913a+/eve6y7du3691339WHH36oDz/8UJ07d9apU6c0YsQIzZgxQ0uXLtX27dv11VdfSZKGDh2qZ555RsuXL5dlWZo3b54/qwwAAAAUya8Bet68eRo9erQiIiIkSVlZWTpw4ICeeeYZJSYmaurUqXK5XNq2bZvq16+vevXqyeFwKDExUcuWLdP+/ft16tQpxcbGSpJuv/12LVu2zJ9VBgAAAIrk8OfBX3zxRY/XR44cUbt27TRmzBhVqVJFjzzyiObPn68qVaooPDzcvV1ERIRSUlKUmprqUR4eHq6UlBSjOtSuXfXCLqIUCQ4OUnDwH78yhyPIa5ndJrlsNnd5YdspJ++cxwsODjpHGwZJksLDq13o5Z2X4GD/nj9Q11VW0D6Fo22KRvsUjfYpGu1TONqmZPg1QJ+tXr16eu2119yv+/Xrp6SkJHXt2rXAtjabTZZleS03ceRIhlyugscpa+x2m5zOPDmdf4wBz83N81rmsiTLstzlhW0n6ZzHczqDimzD6tVPHyct7UQxXKU5p/P0+Q8dKv7zh4dX88txywvap3C0TdFon6LRPkWjfQpH2xStOD9clOgsHMnJyVq+fLn7tWVZcjgcioyM1OHDh93lqampioiIKFB+6NAh93AQAAAAIBBKNEBblqWxY8cqLS1NTqdTH3zwgTp37qxmzZppz549+vnnn5WXl6clS5YoLi5OUVFRqlSpkjZt2iRJSkpKUlxcXElWGQAAAPBQokM4GjdurIcfflh33323cnNzFR8fr27dukmSxo8fr0GDBik7O1sdOnRwD+uYOHGiRo4cqczMTDVp0kT9+/cvySoDAAAAHkokQH/++efun/v27au+ffsW2KZ9+/b66KOPCpQ3btxY8+fP92v9AAAAAF+xEiEAAABggAANAAAAGCBAAwAAAAYI0AAAAICBEp2FA2WT3V5w8ZrysDgNAADA+SBAo0hhoSGaszxZ6Zk5HmV9O0cTogEAQIVEgMY5pWfmKC0jO9DVAAAAKBUYAw0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAhVRwXvKX97bZ/njNyoQAAKAioAcaxvKX956ZtF0HDmfqwOFMzf10pztUAwAAlGf0QOO85C/vnZvncr8GAACoCOiBBgAAAAwQoAEAAAADBGgAAADAgE8BesSIEQXKBg0aVOyVAQAAAEq7Ih8iHD16tFJSUrRp0yYdPXrUXZ6bm6vdu3f7vXIAAABAaVNkgO7du7d27dql5ORkdenSxV0eFBSk5s2b+71yAAAAQGlTZIC+5pprdM011+i6667TJZdcUlJ1AgAAAEotn+aB3rdvn4YOHaq0tDRZ1h+rzS1evNhvFQMAAABKI58C9JgxY9SrVy81adJENhurzQEAAKDi8ilABwcH64EHHvB3XQAAAIBSz6dp7Bo1aqTk5GR/1wUAAAAo9Xzqgf7ll1/Uq1cvXXrppapUqZK7nDHQAAAAqGh8CtBDhgzxdz0AAACAMsGnAB0dHe3vegAAAABlgk8Bul27drLZbLIsyz0LR3h4uFauXOnXygEAAACljU8B+scff3T/7HQ69cknn3iUAQAAABWFT7NwnCk4OFgJCQlavXq1P+oDAAAAlGo+9UAfP37c/bNlWdq+fbvS09P9VScAAACg1DIeAy1JtWvX1tNPP+3XigEAAAClkfEYaAAAAKAi8ylAu1wuzZ49WytXrlRubq6uv/56Pfroo3I4fNodAAAAKDd8eojwlVde0dq1a3XffffpgQce0LfffqsJEyb4u24AAABAqeNTF/KqVau0YMECBQcHS5Juuukmde/eXSNGjPBr5QAAAIDSxqceaMuy3OFZkkJCQjxeAwAAABWFTwG6cePGGjt2rPbt26d9+/Zp7NixLO8NAACACsmnAD169Gilp6erT58+uvPOO3Xs2DE988wz/q4bAAAAUOoUGaBzcnI0bNgwrV27VuPHj9eaNWt07bXXKigoSFWrVi2pOgIAAAClRpEBeurUqcrIyFDz5s3dZc8//7zS09M1bdo0v1cOAAAAKG2KDNBffvmlXnnlFdWuXdtdFhkZqQkTJuizzz7ze+UAAACA0qbIAB0cHKzKlSsXKK9atapCQkL8VikAAACgtCoyQNvtdmVkZBQoz8jIUG5urt8qBQAAAJRWRQbobt26aeTIkcrKynKXZWVlaeTIkYqPj/d75QAAAIDSpsgAfd9996latWq6/vrrdeedd6p37966/vrrFRYWpgEDBpRUHQEAAIBSo8ilvO12u55//nk98sgj+uGHH2S323XNNdcoMjKypOoHAAAAlCpFBuh8devWVd26df1dFwAAAKDU82klQgAAAACnEaABAAAAAwRoAAAAwAABGgAAADBAgAYAAAAMEKABAAAAAwRoAAAAwIBP80ADvrDbbQXKXC4rADUBAADwHwI0ikVYaIjmLE9WemaOR1nfztGEaAAAUK4QoFFs0jNzlJaRHehqAAAA+BVjoAEAAAADBGgAAADAAAEaAAAAMECABgAAAAwQoAEAAAADBGgAAADAAAEaAAAAMECABgAAAAwQoAEAAAADBGgAAADAAAEaAAAAMECABgAAAAw4Al0BlG92u61AmctlBaAmAAAAxYMADb8JCw3RnOXJSs/M8Sjr2zmaEA0AAMosAjT8Kj0zR2kZ2YGuBgAAQLFhDDQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABvwbojIwMdevWTb/++qskac2aNUpMTFR8fLwmT57s3m7Hjh3q1auXunTpoqefflq5ubmSpAMHDqhv377q2rWrHnvsMWVmZvqzugAAAMA5+S1Ab926VXfffbf27t0rSTp16pRGjBihGTNmaOnSpdq+fbu++uorSdLQoUP1zDPPaPny5bIsS/PmzZMkPffcc7rnnnu0bNkyNW3aVDNmzPBXdQEAAACf+C1Az5s3T6NHj1ZERIQkadu2bapfv77q1asnh8OhxMRELVu2TPv379epU6cUGxsrSbr99tu1bNkyOZ1ObdiwQV26dPEoBwAAAALJ4a8Dv/jiix6vU1NTFR4e7n4dERGhlJSUAuXh4eFKSUnRsWPHVLVqVTkcDo9ylH12u61AmctlBaAmAAAA5vwWoM9mWQUDks1mMy43Vbt2VeN9Sqvg4CAFB//xK3M4gryW2W2Sy2Zzlxe2nXLyfDpeUWX5vxNf961V/SL9e8X/dCIrx11WrUqI/tK72Xm3iSSFh1c7r/3PxV/HLS9on8LRNkWjfYpG+xSN9ikcbVMySixAR0ZG6vDhw+7XqampioiIKFB+6NAhRUREqFatWsrIyFBeXp6CgoLc5aaOHMkoF72bdrtNTmeenM5cd1lubp7XMpd1+gNLfnlh20ny6XhFleV/0DHZNz0zR2kZ2e4ypzPvvH9PTufp6zh06ITxvucSHl7NL8ctL2ifwtE2RaN9ikb7FI32KRxtU7Ti/HBRYtPYNWvWTHv27NHPP/+svLw8LVmyRHFxcYqKilKlSpW0adMmSVJSUpLi4uIUHBysVq1aaenSpR7lAAAAQCCVWA90pUqVNH78eA0aNEjZ2dnq0KGDunbtKkmaOHGiRo4cqczMTDVp0kT9+/eXJI0ePVrDhw/XzJkzVadOHU2aNKmkqgsAAAB45fcA/fnnn7t/bt++vT766KMC2zRu3Fjz588vUB4VFaU5c+b4tX4AAACACVYiBAAAAAwQoAEAAAADJTYGGigKc0MDAICyggCNgAsLDdGc5clKz8zxKOvbOZoQDQAASh0CNEqFs+eGBgAAKK0YAw0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABgjQAAAAgAFHoCsA7+x2W5GvAQAAEBgE6FLIbrdp7qc7lZ6Z4y67NDw0gDUCAABAPgJ0KZWemaO0jGz367DQkADWBgAAAPkYAw0AAAAYIEADAAAABgjQAAAAgAECNAAAAGCAAA0AAAAYIEADAAAABpjGDqWWt8VjXC4rADUBAAD4AwEapVJYaIjmLE/2WEwmLDREfTtHE6IBAEBAEaBRap29mAwAAEBpwBhoAAAAwAABGgAAADBAgAYAAAAMEKABAAAAAwRoAAAAwAABGgAAADBAgAYAAAAMEKABAAAAAwRoAAAAwAABGgAAADBAgAYAAAAMEKABAAAAAwRoAAAAwAABGgAAADBAgAYAAAAMEKABAAAAAwRoAAAAwAABGgAAADBAgAYAAAAMEKABAAAAA45AVwAwYbfbAl0FAABQwRGgUWaEhYZozvJkpWfmuMtSj51URM2LAlgrAABQ0RCgUaakZ+YoLSPb/To3zxXA2gAAgIqIMdAAAACAAQI0AAAAYIAADQAAABggQAMAAAAGCNAAAACAAQI0AAAAYIBp7IqJtwU+XC4rADUBAACAPxGgi4HdbtPcT3d6LPARFhqivp2jCdEAAADlDAG6mJy9wAcAAADKJ8ZAAwAAAAbogfYjxkUDAACUPwRoPwkLDdGc5cmMiwYAAChnCNB+xLhoAACA8ocx0AAAAIABeqBLmLdx0eezDQAAAAKDAF2CvI2LvjQ8VBlZzgJlAAAAKJ0I0CXs7HHRYaEhXssAAABQOjEGGgAAADBAgAYAAAAMEKABAAAAA4yBPg9nz5LBrBkAAAAVBwHakN1u09xPdzJrBgAAQAVFgD4PzJoBAABQcTEGGgAAADBAgAYAAAAMEKABAAAAAwRoAAAAwAABGgAAADBAgAYAAAAMEKABAAAAAwRoAAAAwAALqaDc8ba0ustlBaAmAACgPCJAo8yz2f4IzXa7TXOWJ3sstR4WGqK+naMJ0QAAoFgQoFGmOYLsSjl6UjOTtkuSLg0PLbDUOgAAQHEiQKPMy81zuQNzWGhIgGsDAADKOx4iBAAAAAwQoAEAAAADBGgAAADAAAEaAAAAMECABgAAAAwQoAEAAAADBGgAAADAAAEaAAAAMECABgAAAAywEiEqLLvdVqDM5bICUBMAAFCWEKBRIdntNs39dKfSM3PcZWGhIerbOZoQDQAAikSARoWVnpmjtIzsQFcDAACUMYyBBgAAAAwQoAEAAAADDOEAznD2g4V2u40x0QAAwAMBGvhdWGiI5ixPdj9YGBwcpItCgniwEAAAeCBAo0Lw1rPszZkPFgYHO+SsFOT3ugEAgLKFAI1y7+yeZUm6NDw0gDUCAABlGQEaFcLZU9aFhYYEsDYAAKAsYxYOAAAAwAABGgAAADBAgAYAAAAMMAYaOAdvM3YwrR0AABUXARoogrcZPMJCQ5gbGgCACowADZzD2TN4AACAio0x0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAYC8hBh//79deTIETkcp08/ZswY7du3TzNnzpTT6dT999+vvn37SpLWrFmjcePGKTs7W3/60580ZMiQQFQZAAAAkBSAAG1Zlnbv3q0vv/zSHaBTUlI0ZMgQLVy4UCEhIerTp4/atm2runXrasSIEZozZ47q1KmjRx55RF999ZU6dOhQ0tUGAAAAJAUgQO/evVs2m00PPfSQjhw5ojvvvFOhoaFq166datSoIUnq0qWLli1bpjZt2qh+/fqqV6+eJCkxMVHLli0jQAMAACBgSnwMdHp6utq3b6/XXntNb7/9tt5//30dOHBA4eHh7m0iIiKUkpKi1NRUr+UAAABAoJR4D3Tz5s3VvHlzSVKVKlXUu3dvjRs3To8++qjHdjabTZZVcKU3m63gsspFqV276vlXthDBwUEKDv6j6RyOoFJVZrdJLpvNXV7YdsrJu+Dz5v8+AnW9thzJZj/3tRZnWXBwkF/uq7IoPLxaoKtQatE2RaN9ikb7FI32KRxtUzJKPEBv3LhRTqdT7du3l3R6THRUVJQOHz7s3iY1NVURERGKjIz0Wm7iyJGMYl1y2W63yenMk9OZ6y7Lzc0rVWUu63S75pcXtp2kCz5v/oecQF2vZUmW69zXej5lwcEOr9s5nUHFfl+VReHh1XTo0IlAV6NUom2KRvsUjfYpGu1TONqmaMX54aLEh3CcOHFCEyZMUHZ2tjIyMrRo0SK9/PLL+uabb3T06FGdPHlSn3zyieLi4tSsWTPt2bNHP//8s/Ly8rRkyRLFxcWVdJUBAAAAtxLvge7YsaO2bt2q2267TS6XS/fcc49atmypIUOGqH///nI6nerdu7euvfZaSdL48eM1aNAgZWdnq0OHDuratWtJVxkAAABwC8g80E888YSeeOIJj7LExEQlJiYW2LZ9+/b66KOPSqhmAAAAQNFYiRAAAAAwEJAeaKCss9sLzgZT0R8qBACgoiBAA4bCQkM0Z3my0jNzPMr6do4mRAMAUAEQoIHzkJ6Zo7SM7EBXAwAABAABGigmDOsAAKBiIEADxYBhHQAAVBwEaKCYMKwDAICKgWnsAAAAAAMEaAAAAMAAARoAAAAwQIAGAAAADBCgAQAAAAMEaAAAAMAAARoAAAAwwDzQPjhzhTlvq80BJlixEACAso0AfQ52u01zP93pXmHu0vDQANcIZdnZ95PEioUAAJQ1BGgfnLnCXFhoSIBrg7KOFQsBACjbGAMNAAAAGCBAAwAAAAYYwgH40dkPDPIQKgAAZR8BGvCTsNAQzVme7PHAIA+hAgBQ9hGgAT86+4FBHkIFAKDsI0ADpRTzRQMAUDoRoIFSiPmiAQAovQjQQCnFfNEAAJROBGigFGC2DgAAyg4CNBBgzNYBAEDZQoAGSgFm6wAAoOxgJUIAAADAAAEaAAAAMECABgAAAAwQoAEAAAADBGgAAADAAAEaAAAAMECABgAAAAwwDzRQhnhbodDlsgJQEwAAKi4CNFBGeFuxMCw0RP26xBQI0YRqAAD8hwANlCHeViz0Fqr7do4mRAMA4CcEaKCMOztUAwAA/+IhQgAAAMAAPdBABcEDiAAAFA8CNFAB2O02zf10J2OlAQAoBgRooIJgrDQAAMWDMdAAAACAAQI0AAAAYIAhHEA5dPYDg94eIAQAAOeHAA2UM94WV7k0PDSANQIAoHwhQAPlkLcVCwEAQPFgDDQAAABggAANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAAAAGCBAAwAAAAaYBxqowHxdodDlsvxcEwAAyg4CNFBBFbZiYUaW06MsLDREfTtHew3RZwdwgjYAoCIgQAMVmLcVC88u88Zut2nG/K06mnbSY9/CgjYAAOUJARrAeTmRde6gDQBAecRDhAAAAIABAjQAAABggAANAAAAGCBAAwAAAAZ4iBBAsfE2rzSzcgAAyhsCNIBi4W1eaaa2AwCURwRoAOd0ds9yYSsYeptDml5pAEB5Q4AGUKTCViw8333plQYAlHUEaADn5G3FwvPdFwCAso5ZOAAAAAADBGgAAADAAEM4AAQcDxoCAMoSAjSAgLLbbZr76U4eNAQAlBkEaAABx4OGAICyhAANoMSdOWSjsDmlAQAorQjQAErU2XND+zqntMRYaQBA6UCABlDizhyy4euc0oyVBgCUFgRoAKWSt+XDWSocAFAaEKABlDq+Lh/OUuEAgEAgQAMolXxdPpwZPAAAJY0ADaDcYVgHAMCfCNAAyhWGdQAA/I0ADaDc4WFDAIA/EaABlHv0SgMAihMBGkCF4OvDhvRUAwDOhQANAL8rbLGWfl1iCoRoQjUAVFwEaAAVli+LtTD8AwBwNgI0gArJ18VaJB5KBAB4IkCfxVuPFIDyydfFWs5GrzQAVGwE6DN4G/9YWI8UgIqNFRABoOIiQJ/lfHukAODMb6zyf6ZHGgDKHwI0ABSDM4d1BAcHyenMYwYPACinCNAAUEzyv8EKDnbI6cw1Givt60OJPLwIAIFHgAYAP/JlBg+73eZT0C5snmoeXgSAkkWABoASVNj0eb4+lMjDiwAQeARoAChhJfGwMkM9AMB/CNAAUMb5OiSEBxoBoHgQoAGglPJlYSdfh4SUtsVfzr4WgjyAsoQADQCl0IUsNV7YkJDSMH7abrdpxvytOpp20l3Gg5AAyhoCNACUUoFa2Mnf46dPZAU+yAPAhSBAAwDcmCoPAM6NAA0A8FAahnoAQGlGgAYAnJMvD/0xdR6AioIADQAV2PnM9OFtSjxfV1Ms7vpKhHQAJY8ADQAV1PnO9HGhqyl640swLmx8tq/zWxO+ARQXAjQAVGDnO9OHr/ud71zWhfVe+zq/daB6yAFUDARoAIBfXMhc1pJv4dvbviY95PRKAzgfBGgAgN+cbw+3Sfg+3/P6Y3VGAjlQMRCgAQClUkksJFOcU/aZzKFN0AbKNgI0AADFxJdA7o/FaoozkPt6LD4EoCKr0AHa1/F1AICK43z/LyhsP2//15xrPHb+z74E18IekDz7QUpfwq2vM53wUCYqugoboL39I2Eyvg4AUP4UNvY6I8vpU5mvxytqu+DgIDmdeT4HV28PSPoyd7c33sK9P6YtBMq6ChugpZIZXwcAKFu8/d/ga5mvxytqu+Bgh5zOXJ+Dqy/nNflgcCHX4M2FDAnx9Zj0eqOkVegADQBAaVacHT0X8iHgfF3IkJDCAv6pHJeOpp30OB4PaqKkEaABAECx8GW8t0nPureykzl5Ps0Z7usY7ZJ4aNLX3nUCftlBgAYAABfsQhbOuZCe7wtZOMfXBzAvJJD72rvu6zh1QnbpUCYC9OLFizVz5kw5nU7df//96tu3r/ExmHEDAAD/CtSzRee7cI4vD2AWtp3ke1j2tXf9zH1NQvaF9IQTyM9PqQ/QKSkpmjx5shYuXKiQkBD16dNHbdu21ZVXXunzMZhxAwAAFOeDnxf6kOe5zutLyM4vOzNUFzWdoi+97SWlrM8tbrMsq3TW7HeLFi3Shg0bNHbsWEnSa6+9JsuyNHDgQJ/2P3YsU5KU9PUeZWQ53eWRtS5S5qlc47Lz3a8kyyJXzFNunvRDq8Qit8t2unT8RPZ5nSO/rGfHkZKkNVtfCcj1JqWPVl6epV41xxT78R0Ou2qHVSpVv9vSVFYc90+gr8FfZQ6HXbm5rlJRF3+UlZZ7p7yW+eP+KQ3XVVxlFeX+MdnPbrcp82Su+965uEZlnczOVebJXPd2F9eorNRjJwvd98ztvO1bnGWhFznUuVW9AsNkPt34yzm3uxA1axZf52mpD9Cvv/66srKyNGTIEEnSf/7zH23btk3PP/98gGsGAACAisge6Aqci7d8b7MxfhkAAACBUeoDdGRkpA4fPux+nZqaqoiIiADWCAAAABVZqQ/Q1113nb755hsdPXpUJ0+e1CeffKK4uLhAVwsAAAAVVKmfhSMyMlJDhgxR//795XQ61bt3b1177bWBrhYAAAAqqFL/ECEAAABQmpT6IRwAAABAaUKABgAAAAwQoAEAAAADBGgAAADAQLkN0IsXL9att96qzp07a+7cuYGuTkD0799fCQkJ6tGjh3r06KGtW7cW2i5r1qxRYmKi4uPjNXny5ADW2v8yMjLUrVs3/frrr5IKv/YdO3aoV69e6tKli55++mnl5p5eXvTAgQPq27evunbtqscee0yZmZkBuQ5/Obt9nnrqKcXHx7vvo08//VSSebuVddOnT1dCQoISEhI0YcIESdw7Z/LWPtw7f5gyZYpuvfVWJSQk6K233pLE/XMmb+3D/ePppZde0vDhwyWZ3yPp6el6+OGH9ac//Ul9+/bVoUOHAnYd/nJm+0yfPl0dO3Z03zv5eadY/7ascui3336zOnbsaB07dszKzMy0EhMTrV27dgW6WiXK5XJZ119/veV0Ot1lhbXLyZMnrQ4dOlj79u2znE6n9eCDD1pffvllAGvvP1u2bLG6detmXX311dYvv/xS5LUnJCRY3377rWVZlvXUU09Zc+fOtSzLsh5++GFryZIllmVZ1vTp060JEyYE5Fr84ez2sSzL6tatm5WSkuKx3fm0W1m2evVq66677rKys7OtnJwcq3///tbixYu5d37nrX0++eQT7p3frVu3zurTp4/ldDqtkydPWh07drR27NjB/fM7b+3z008/cf+cYc2aNVbbtm2tYcOGWZZlfo8899xz1uuvv25ZlmUtWrTIevzxx0v2Avzs7PZ55JFHrM2bNxfYrjj/tsplD/SaNWvUrl071ahRQ1WqVFGXLl20bNmyQFerRO3evVs2m00PPfSQunfvrnfffbfQdtm2bZvq16+vevXqyeFwKDExsdy217x58zR69Gj3apaFXfv+/ft16tQpxcbGSpJuv/12LVu2TE6nUxs2bFCXLl08ysuLs9snKytLBw4c0DPPPKPExERNnTpVLpfLuN3KuvDwcA0fPlwhISEKDg5Ww4YNtXfvXu6d33lrnwMHDnDv/K5NmzZ655135HA4dOTIEeXl5Sk9PZ3753fe2qdSpUrcP787fvy4Jk+erEcffVSSzuse+fLLL5WYmChJ6tatm1auXCmn01nyF+MHZ7ePJG3fvl2zZs1SYmKixowZo+zs7GL/2yqXATo1NVXh4eHu1xEREUpJSQlgjUpeenq62rdvr9dee01vv/223n//fR04cMBru1Sk9nrxxRfVqlUr9+vCrv3s8vDwcKWkpOjYsWOqWrWqHA6HR3l5cXb7HDlyRO3atdPYsWM1b948bdy4UfPnzzdut7KuUaNG7n909+7dq6VLl8pms3Hv/M5b+9x4443cO2cIDg7W1KlTlZCQoPbt2/Nvz1nObp+8vDzun9+NGjVKQ4YMUVhYmKSC/2/5co+cuY/D4VDVqlV19OjREr4S/zi7fTIzM3XVVVdp2LBhWrRokdLT0zVjxoxi/9sqlwHa8rI2jM1mC0BNAqd58+aaMGGCqlSpolq1aql3796aOnVqge1sNluFbq/Crt20vLyqV6+eXnvtNdWuXVsXXXSR+vXrp6+++qrCts+uXbv04IMPatiwYbrssssKvF/R750z26dBgwbcO2cZPHiwvvnmGx08eFB79+4t8H5Fv3/ObJ9vvvmG+0fSf/7zH9WpU0ft27d3lxVXG9jtZT8Cemuf0NBQzZo1S/Xr15fD4dCDDz7ol3un1C/lfT4iIyO1ceNG9+vU1FT3V9IVxcaNG+V0Ot03lWVZioqK0uHDh93b5LdLZGSk1/KKoLBrP7v80KFDioiIUK1atZSRkaG8vDwFBQW5y8ur5ORk7d271/3VlmVZcjgcxu1WHmzatEmDBw/WiBEjlJCQoPXr13PvnOHs9uHe+cNPP/2knJwcXXXVVbrooosUHx+vZcuWKSgoyL1NRb5/vLXP0qVLVaNGjQp//yxdulSHDh1Sjx49lJaWpqysLNlsNuN7JCIiQocPH9Yll1yi3NxcZWRkqEaNGgG6quLjrX2eeuoptWzZUr1795ZU+L1zoX9bZf/jhxfXXXedvvnmGx09elQnT57UJ598ori4uEBXq0SdOHFCEyZMUHZ2tjIyMrRo0SK9/PLLXtulWbNm2rNnj37++Wfl5eVpyZIlFaa9Crv2qKgoVapUSZs2bZIkJSUlKS4uTsHBwWrVqpWWLl3qUV5eWZalsWPHKi0tTU6nUx988IE6d+5s3G5l3cGDBzVgwABNnDhRCQkJkrh3zuStfbh3/vDrr79q5MiRysnJUU5OjlasWKE+ffpw//zOW/u0bt2a+0fSW2+9pSVLlujDDz/U4MGD1alTJ40bN874HunQoYOSkpIknQ6drVq1UnBwcECuqTh5a5+hQ4fq5Zdf1i+//CLLsjR37lx17ty52P+2bJa3vutyYPHixXr99dfldDrVu3dvPfTQQ4GuUol79dVXtXz5crlcLt1zzz267777Cm2Xb775RuPGjVN2drY6dOigp556qlx8/VWYTp066Z133lHdunULvfYff/xRI0eOVGZmppo0aaJx48YpJCRE+/fv1/Dhw3XkyBHVqVNHkyZNUvXq1QN9ScXqzPaZO3eu5s6dq9zcXMXHx+tvf/ubpMLvmcLarSx74YUXtGDBAo9hG3369NHll1/OvaPC28flclX4eyff1KlT3b3O8fHxGjRoEP/2nMFb+/Bvj6eFCxdq/fr1Gj9+vPE9cvz4cQ0fPly//PKLqlWrpokTJ6pu3bqBvqRidWb7LF++XNOmTZPT6VSLFi303HPPKSQkpFj/tsptgAYAAAD8oVwO4QAAAAD8hQANAAAAGCBAAwAAAAYI0AAAAIABAjQAAABggAANAKXMv//9b73xxht+PUenTp303XffFftx161bp27dunl9b9u2bRo1alSxnxMASlq5XIkQAMqyu+++O9BV8Iv//e9/SklJCXQ1AOCCEaABoJitW7dOL774oqpUqaKsrCwNHjzYvYBR5cqVNWzYMDVr1kwdO3bU9OnTdc0110iShgwZotatW+vIkSM6duyYRo0apZSUFI0ZM0YHDx6U0+lUQkKCHn30UQ0YMEA33XST7rjjDm3ZskV33XWXPvvsM9WrV08zZ87UiRMn9Pe//92n+n7++eeaOXOmR/2aN2+uadOmaf/+/Tp06JD279+vWrVqafLkyYqMjNS2bdv07LPPyul06rLLLtOBAwc0fPhwSVJWVpaGDBmi3bt3Kzs7Wy+88IKioqI0depUnThxQk899ZTGjRvnt/YHAH9jCAcA+MGuXbv0yiuvaOrUqZoyZYreeOMNJSUl6fnnn9egQYN06tQp9erVS4sWLZIkpaWlac2aNUpMTPQ4ztChQ9WrVy8tXLhQ8+fP15o1a7R06VJ17txZq1atkiStWrVK4eHhWrNmjSRpxYoV6tq1q0/13Lt3ryZPnlygfllZWZKkjRs3asqUKVq2bJnCwsL0wQcfKDc3V4MGDdLjjz+uxYsXq1+/ftqxY4f7mL/99pvuv/9+ffjhh+rTp4+mTZumOnXqaPDgwWrVqhXhGUCZRw80APhBnTp1FBUVpblz5yo1NVX333+/+z2bzaZ9+/apV69e6t27t4YPH64lS5aoY8eOqlatmnu7rKwsbdiwQWlpaZoyZYq77Mcff9Sf//xnjRs3Trm5ufr666/12GOPafXq1brpppt05MgRd6/2uaxevbrQ+klSmzZtVLVqVUlSkyZNlJaWpp07d0qSOnToIElq166dGjVq5N6/Xr16atasmSSpcePGWrBggWHrAUDpRoAGAD+oUqWKJMnlcql9+/Z69dVX3e8dPHhQERERCgoKUpMmTfTll19q4cKFGjFihMcxXC6XLMvS+++/r4suukiSdPToUVWqVEmhoaFq0qSJvvjiC504cUI9evTQa6+9ps8++0y33HKLbDabT/Usqn6ffvqpKleu7C632WyyLEtBQUGyLMvjOEFBQe6fg4ODC+wDAOUJQzgAwI/atWun1atX66effpIkffXVV+revbuys7MlSXfeeadmzZqlU6dOqWXLlh77Vq1aVbGxsXrrrbckSenp6br77ru1YsUKSdItt9yiSZMmqX379qpataquuOIKzZo1S126dCm2+nnTsGFDhYSEaOXKlZJOz66xc+fOc4b2oKAg5ebm+lw3ACitCNAA4EeNGjXSmDFj9OSTT6p79+6aMmWKZs6c6e6h7tSpk/bv36/evXt73X/ixInaunWrEhMTdccdd6hbt27q3r27pNMBes+ePbruuuskSTfccINyc3PVokWLYqufNw6HQ9OmTdP06dN122236c0339TFF1/s0VvtTfPmzbV7924NGDDA5/oBQGlks/huDQBg6KWXXtKf//xnXXzxxTp48KB69Oihzz77TGFhYYGuGgD4HWOgAaAc+uijjzR79myv7yUmJur//u//Luj4UVFRuv/+++VwOGRZll544QXCM4AKgx5oAAAAwABjoAEAAAADBGgAAADAAAEaAAAAMECABgAAAAwQoAEAAAADBGgAAADAwP8D43IOXwtIAl4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(rc={'figure.figsize':(20,30)})\n",
    "sns.displot(train_data['review_lenght'], height=10, aspect=1)\n",
    "plt.xlim(0, 4500)\n",
    "plt.axvline(mean, color = 'green', label=\"mean\")\n",
    "plt.axvline(mode, color = 'r', label=\"mode\")\n",
    "plt.axvline(median, color = 'yellow', label=\"median\")\n",
    "plt.title('Words per review distribution', fontsize=20)\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def tfidf_features(X_train, X_val, X_test):\n",
    "    \"\"\"\n",
    "        X_train, X_val, X_test — samples        \n",
    "        return TF-IDF vectorized representation of each sample and vocabulary\n",
    "    \"\"\"\n",
    "    # Create TF-IDF vectorizer with a proper parameters choice\n",
    "    # Fit the vectorizer on the train set\n",
    "    # Transform the train, test, and val sets and return the result\n",
    "    tfidf_vectorizer = TfidfVectorizer(token_pattern='(\\S+)', min_df=5, max_df=0.9, ngram_range=(1,2))\n",
    "    tfidf_vectorizer.fit(X_train)\n",
    "    X_train = tfidf_vectorizer.transform(X_train)\n",
    "    X_val = tfidf_vectorizer.transform(X_val)\n",
    "    X_test = tfidf_vectorizer.transform(X_test)\n",
    "    return X_train, X_val, X_test, tfidf_vectorizer.vocabulary_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "train, test = train_test_split(train_data, test_size=0.15)\n",
    "X_train = train.review\n",
    "X_val = test.review\n",
    "X_test = test_data['review']\n",
    "y_train = train.sentiment.to_numpy()\n",
    "y_val = test.sentiment.to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_tfidf, X_val_tfidf, X_test_tfidf, tfidf_vocab = tfidf_features(X_train, X_val, X_test)\n",
    "tfidf_reversed_vocab = {i:word for word,i in tfidf_vocab.items()}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Naive Bayes classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Naive Bayes Accuracy Score ->  87.99259009767599\n"
     ]
    }
   ],
   "source": [
    "nbc = naive_bayes.MultinomialNB()\n",
    "nbc.fit(X_train_tfidf,y_train)\n",
    "# predict the labels on validation dataset\n",
    "predictions_NB = nbc.predict(X_val_tfidf)\n",
    "# Use accuracy_score function to get the accuracy\n",
    "print(\"Naive Bayes Accuracy Score -> \",accuracy_score(predictions_NB, y_val)*100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Accuracy Score ->  89.47457056247895\n"
     ]
    }
   ],
   "source": [
    "log_reg = LogisticRegression()\n",
    "log_reg.fit(X_train_tfidf,y_train)\n",
    "predictions_log_reg = log_reg.predict(X_val_tfidf)\n",
    "# Use accuracy_score function to get the accuracy\n",
    "print(\"Logistic Regression Accuracy Score -> \",accuracy_score(predictions_log_reg, y_val)*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}