sms_adversarial/original_classification

classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "headers = list(word2vec_df)\n",
    "headers.remove('label_not_spam')\n",
    "headers.remove('label_spam')\n",
    "\n",
    "X = np.array(word2vec_df[headers].values.tolist())\n",
    "y = np.array(word2vec_df[['label_not_spam', 'label_spam']].values.tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Value counts for training \n",
      "\n",
      "3900\n",
      "\n",
      "\n",
      "Value counts for testing \n",
      "\n",
      "1672\n"
     ]
    }
   ],
   "source": [
    "# Split data to training (70%) and testing (30%)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)\n",
    "\n",
    "print(\"Value counts for training \\n\")\n",
    "print(y_train[:, 0].size)\n",
    "print(\"\\n\")\n",
    "print(\"Value counts for testing \\n\")\n",
    "print(y_test[:, 0].size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.97      0.96      0.97      1462\n",
      "           1       0.75      0.82      0.78       210\n",
      "\n",
      "   micro avg       0.94      0.94      0.94      1672\n",
      "   macro avg       0.86      0.89      0.88      1672\n",
      "weighted avg       0.95      0.94      0.94      1672\n",
      " samples avg       0.94      0.94      0.94      1672\n",
      "\n"
     ]
    }
   ],
   "source": [
    "def classify(X_train, y_train, X_test, y_test):\n",
    "    # Initialise Decision Tree\n",
    "    clf = DecisionTreeClassifier()\n",
    "    # Fit model\n",
    "    model = clf.fit(X_train, y_train)\n",
    "    # Predict testing target labels\n",
    "    prediction = model.predict(X_test)\n",
    "    \n",
    "    return prediction\n",
    "\n",
    "print(classification_report(y_test, classify(X_train, y_train, X_test, y_test)))"
   ]
  }
 ],
 "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}