SVM .ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split,cross_val_score\n",
    "from sklearn.svm import LinearSVC,SVC\n",
    "import timeit\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import confusion_matrix,classification_report,accuracy_score,f1_score,precision_score,recall_score,RocCurveDisplay\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_resampled=pd.read_csv('resampled_without_encoding.csv')\n",
    "x_resampled_encoding=pd.read_csv('resampled_with_encoding.csv')\n",
    "y_resampled=pd.read_csv('y_resampled.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data with Resampling and with OneHotEncoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time taken to fit the model in sec is  607.1905270999999\n",
      "Accuracy score: 0.9005\n",
      "F1 score : 0.8974\n",
      "Recall score : 0.8681\n",
      "Precision score : 0.9289\n"
     ]
    }
   ],
   "source": [
    "X = x_resampled_encoding.values\n",
    "Y = np.squeeze(y_resampled.values)\n",
    "X_train,X_test,y_train,y_test=train_test_split(X,Y,test_size=0.2,random_state=10)\n",
    "\n",
    "scaler = StandardScaler().fit(X_train)\n",
    "train_sc = scaler.transform(X_train)\n",
    "test_sc = scaler.transform(X_test)\n",
    "\n",
    "svm_clf = LinearSVC(random_state=48) \n",
    "start_time = timeit.default_timer()\n",
    "svm_clf.fit(train_sc,y_train)\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print(\"Time taken to fit the model in sec is \",elapsed)\n",
    "\n",
    "y_pred=svm_clf.predict(test_sc)\n",
    "print(f\"Accuracy score: {accuracy_score(y_test, y_pred):.4f}\")\n",
    "print(f\"F1 score : {f1_score(y_test, y_pred):.4f}\")\n",
    "print(f\"Recall score : {recall_score(y_test, y_pred):.4f}\")\n",
    "print(f\"Precision score : {precision_score(y_test, y_pred):.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(confusion_matrix(y_test,y_pred),annot=True,fmt='',cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.88      0.93      0.90    114386\n",
      "           1       0.93      0.87      0.90    115022\n",
      "\n",
      "    accuracy                           0.90    229408\n",
      "   macro avg       0.90      0.90      0.90    229408\n",
      "weighted avg       0.90      0.90      0.90    229408\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.roc_curve.RocCurveDisplay at 0x2eb1da15280>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "RocCurveDisplay.from_predictions(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.roc_curve.RocCurveDisplay at 0x2eb7b678340>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "RocCurveDisplay.from_estimator(svm_clf, test_sc, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<BarContainer object of 212 artists>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "from sklearn.feature_selection import SelectPercentile, f_classif\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import svm\n",
    "X_indices = np.arange(X.shape[-1])\n",
    "selector = SelectPercentile(f_classif, percentile=10)\n",
    "selector.fit(X, Y)\n",
    "scores = -np.log10(selector.pvalues_)\n",
    "scores /= scores.max()\n",
    "plt.bar(np.arange(X.shape[-1]) - .45, scores, width=.2,\n",
    "        label=r'Univariate score ($-Log(p_{value})$)', color='g')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4.82311947e-05, 8.93654342e-05, 1.00000000e+00, 8.33075651e-01,\n",
       "       9.13603388e-01, 9.79510496e-01, 2.98247142e-01, 3.09974006e-01,\n",
       "       4.03912714e-01, 3.21945349e-01, 1.71212225e-01, 8.84782652e-05,\n",
       "       8.28897245e-02, 1.39301142e-01, 1.56340398e-01, 1.45072258e-03,\n",
       "       2.06941644e-03, 2.51934503e-03, 1.30676481e-02, 2.29410475e-02,\n",
       "       2.05140337e-04, 2.78931807e-05, 2.16820200e-08, 4.87436935e-05,\n",
       "       1.19027505e-03, 1.16923305e-03, 4.60619068e-04, 5.29764259e-04,\n",
       "       3.59503544e-04, 8.44099329e-04, 3.83980701e-05, 8.33509947e-05,\n",
       "       1.24209843e-03, 3.10275443e-04, 3.82219660e-05, 3.75144161e-02,\n",
       "       9.92955269e-03, 1.06961275e-02, 2.01857857e-02, 1.75195149e-02,\n",
       "       2.95378303e-02, 7.24353218e-03, 3.90008164e-06, 1.06976101e-03,\n",
       "       2.95681510e-03, 2.67689735e-03, 2.67689735e-03, 5.51140299e-03,\n",
       "       1.03580861e-03, 7.17376232e-03, 8.92470700e-04, 5.40435397e-03,\n",
       "       2.20281876e-02, 7.68666409e-03, 4.24684389e-04, 5.02347399e-05,\n",
       "       4.84235526e-04, 8.38192735e-04, 1.24222545e-04, 4.61951081e-04,\n",
       "       5.40334867e-05, 3.90556678e-03, 4.40522849e-03, 4.84677655e-04,\n",
       "       7.62643635e-04, 5.33362054e-03, 5.25656308e-03, 5.25656308e-03,\n",
       "       1.12336512e-02, 1.29299935e-02, 5.50911382e-02, 6.42767960e-04,\n",
       "       1.11260385e-03, 8.82922236e-03, 1.53647759e-03, 6.94327295e-02,\n",
       "       7.67294645e-04, 4.00641960e-06, 9.20601108e-03, 2.69265327e-02,\n",
       "       1.59688870e-02, 2.28752473e-04, 2.74217426e-02, 5.05650932e-03,\n",
       "       2.00919479e-02, 1.71221494e-02, 3.23337960e-02, 3.84992713e-04,\n",
       "       1.13236936e-02, 1.99956971e-02, 3.11431546e-02, 7.50749101e-03,\n",
       "       1.13818798e-02, 2.53867720e-03, 1.08248545e-02, 1.27861478e-04,\n",
       "       1.66491344e-03, 1.66491344e-03, 3.02109245e-03, 3.02109245e-03,\n",
       "       4.27675008e-04, 5.03052689e-02, 1.13175041e-03, 2.35346287e-03,\n",
       "       3.56216749e-02, 1.15931009e-02, 1.13840678e-03, 1.99972257e-02,\n",
       "       7.29699291e-02, 1.04238289e-03, 9.33198080e-03, 9.25782034e-03,\n",
       "       1.90741843e-02, 2.31972654e-02, 5.34385556e-05, 2.24730794e-03,\n",
       "       1.78147211e-03, 4.60139430e-03, 6.48573842e-05, 1.42385235e-02,\n",
       "       1.28227928e-04, 6.68972155e-02, 1.67340321e-02, 2.12690351e-02,\n",
       "       4.81396045e-05, 3.80341411e-02, 6.44855090e-03, 3.58849206e-03,\n",
       "       3.86671370e-05, 1.11949949e-02, 2.71834746e-02, 2.61329010e-02,\n",
       "       2.38587211e-04, 2.60728044e-02, 2.72109544e-03, 3.41355885e-02,\n",
       "       1.72056259e-04, 1.09283862e-02, 1.73716332e-02, 4.86168332e-02,\n",
       "       4.26448964e-03, 7.29432697e-02, 2.59264153e-02, 1.13314995e-03,\n",
       "       3.34883890e-02, 2.54164419e-03, 3.88121306e-02, 1.53724371e-03,\n",
       "       3.49208662e-02, 1.31934148e-03, 3.65776255e-02, 5.79903030e-04,\n",
       "       2.12684561e-02, 2.04904207e-02, 3.71602009e-03, 1.91191656e-03,\n",
       "       4.07786832e-03, 3.76205818e-03, 3.73230032e-07, 2.00297993e-04,\n",
       "       1.85622347e-03, 3.61168600e-02, 6.87471683e-02, 1.26696246e-02,\n",
       "       7.28832450e-04, 2.85781054e-02, 1.64393699e-02, 1.34062298e-03,\n",
       "       1.79229602e-02, 1.66675827e-02, 1.39994418e-04, 1.78857385e-02,\n",
       "       5.49201593e-05, 9.03319790e-02, 3.18897750e-03, 3.81512975e-03,\n",
       "       9.11264307e-03, 7.71251718e-03, 2.34381410e-02, 4.68932351e-03,\n",
       "       8.45749492e-03, 5.62773387e-03, 7.61467479e-03, 9.28957229e-04,\n",
       "       6.29075546e-05, 2.79224843e-02, 7.14456109e-03, 2.12436214e-03,\n",
       "       1.04010458e-02, 4.04539644e-03, 7.49890461e-02, 3.95124677e-03,\n",
       "       3.15657325e-03, 5.41718614e-03, 2.58150133e-02, 6.91812381e-02,\n",
       "       6.08737192e-02, 8.03632782e-02, 4.61244615e-04, 7.76013096e-03,\n",
       "       3.58966196e-03, 6.05141094e-02, 5.95688769e-02, 1.03035792e-03,\n",
       "       2.76938012e-04, 2.42583259e-02, 3.46125492e-04, 5.03913585e-03,\n",
       "       8.45042918e-03, 2.95033271e-02, 6.21364423e-03, 5.15910401e-05])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# Compare to the weights of an SVM\n",
    "clf = svm.SVC(kernel='linear')\n",
    "clf.fit(X, Y)\n",
    "\n",
    "svm_weights = (clf.coef_ ** 2).sum(axis=0)\n",
    "svm_weights /= svm_weights.max()\n",
    "svm_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<BarContainer object of 212 artists>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAANzElEQVR4nO3dXYxc513H8e8Pu6mABlLwNgq2gw0yL75oQljcShUlqKK1c2MqgeQUNSWqZCLFqFwgxXBBK/WGIhWhqmksU6y2EtSq1JS6yCRwAeSiBLxGeXMih8UJ8dZRvSGovFQicvLnYibNZDI7c+yMd3af/X6klXfOefbsM88efzU+9hynqpAkrX/fN+sJSJKmw6BLUiMMuiQ1wqBLUiMMuiQ1YvOsvvGWLVtqx44ds/r2krQunT59+oWqmhu1b2ZB37FjBwsLC7P69pK0LiX595X2eclFkhph0CWpEQZdkhph0CWpEQZdkhph0CWpERODnuRYkotJnlhhf5J8JslikseS3DL9aUqSJunyCv0LwN4x+/cBu/ofB4H73vy0JEmXa2LQq+oh4MUxQ/YDX6qeh4HrktwwrQlKkrqZxjX0rcD5gcdL/W1vkORgkoUkC8vLy1P41kAyneNI0jo3jaCPKurI/wapqo5W1XxVzc/NjbwVgSTpCk0j6EvA9oHH24ALUziuJOkyTCPoJ4A7+v/a5d3Ad6rq+SkcV5J0GSbebTHJl4FbgS1JloCPA28BqKojwEngNmAR+C5w59WarCRpZRODXlW3T9hfwN1Tm5Ek6Yr4TlFJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGGHRJaoRBl6RGdAp6kr1JziZZTHJ4xP4fTvKNJI8mOZPkzulPVZI0zsSgJ9kE3AvsA3YDtyfZPTTsbuDJqroJuBX4dJJrpjxXSdIYXV6h7wEWq+pcVb0EHAf2D40p4NokAd4GvAhcmupMJUljdQn6VuD8wOOl/rZBnwV+FrgAPA58rKpeGT5QkoNJFpIsLC8vX+GUJUmjdAl6RmyroccfAB4Bfgy4Gfhskh96wxdVHa2q+aqan5ubu8ypSpLG6RL0JWD7wONt9F6JD7oTuL96FoFngJ+ZzhQlSV10CfopYFeSnf2/6DwAnBga8xzwPoAk1wM/DZyb5kQlSeNtnjSgqi4lOQQ8CGwCjlXVmSR39fcfAT4JfCHJ4/Qu0dxTVS9cxXlLkoZMDDpAVZ0ETg5tOzLw+QXg/dOdmiTpcvhOUUlqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEYYdElqhEGXpEZ0CnqSvUnOJllMcniFMbcmeSTJmST/MN1pSpIm2TxpQJJNwL3ArwBLwKkkJ6rqyYEx1wGfA/ZW1XNJ3nGV5itJWkGXV+h7gMWqOldVLwHHgf1DYz4E3F9VzwFU1cXpTlOSNEmXoG8Fzg88XupvG/RTwNuT/H2S00numNYEJUndTLzkAmTEthpxnJ8H3gd8P/CPSR6uqqdfd6DkIHAQ4MYbb7z82UqSVtTlFfoSsH3g8TbgwogxD1TV/1bVC8BDwE3DB6qqo1U1X1Xzc3NzVzpnSdIIXYJ+CtiVZGeSa4ADwImhMV8HfjHJ5iQ/ALwLeGq6U5UkjTPxkktVXUpyCHgQ2AQcq6ozSe7q7z9SVU8leQB4DHgF+HxVPXE1Jy5Jer1UDV8OXx3z8/O1sLDw5g+UwIyegySttiSnq2p+1D7fKSpJjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjTDoktQIgy5JjegU9CR7k5xNspjk8Jhxv5Dk5SS/Nr0pSpK6mBj0JJuAe4F9wG7g9iS7Vxj3KeDBaU9SkjRZl1foe4DFqjpXVS8Bx4H9I8b9NvBV4OIU5ydJ6qhL0LcC5wceL/W3fU+SrcAHgSPjDpTkYJKFJAvLy8uXO1dJ0hhdgp4R22ro8Z8A91TVy+MOVFVHq2q+qubn5uY6TlGS1MXmDmOWgO0Dj7cBF4bGzAPHkwBsAW5Lcqmq/nIak5QkTdYl6KeAXUl2At8CDgAfGhxQVTtf/TzJF4C/MuaStLomBr2qLiU5RO9fr2wCjlXVmSR39fePvW4uSVodXV6hU1UngZND20aGvKp+881PS5J0uXynqCQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiM2TtAz6i7AktSOjRN0SWqcQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRhh0SWpEp6An2ZvkbJLFJIdH7P+NJI/1P76Z5KbpT1WSNM7EoCfZBNwL7AN2A7cn2T007Bngl6rqncAngaPTnqgkabwur9D3AItVda6qXgKOA/sHB1TVN6vqP/sPHwa2TXeakqRJugR9K3B+4PFSf9tKPgr89agdSQ4mWUiysLy83H2WkqSJugQ9I7bVyIHJL9ML+j2j9lfV0aqar6r5ubm57rOUJE20ucOYJWD7wONtwIXhQUneCXwe2FdV/zGd6UmSuuryCv0UsCvJziTXAAeAE4MDktwI3A98uKqenv40JUmTTHyFXlWXkhwCHgQ2Aceq6kySu/r7jwB/APwo8LkkAJeqav7qTVuSNCxVIy+HX3Xz8/O1sLDw5g+UQJfn0HWcJK1hSU6v9ILZd4pKUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiQ1wqBLUiMMuiSN0/uP79cFgy5pY1pHoe7KoEvSsHUae4MuSY0w6JLUCIOu6Vinf0SVWmLQJakRBl2SGmHQJakRBl16lX8PMB2u48wYdK19BkLqxKBrdgy1NFUGvXVG8+pqfX1bf36NMeham1oOydV8bi2vmybqFPQke5OcTbKY5PCI/Unymf7+x5LcMv2pzoC/OcabxfpspJ9JC8+1heewjkwMepJNwL3APmA3cHuS3UPD9gG7+h8HgfumPM+2edKvL/68ZudK1n7waxr/2XV5hb4HWKyqc1X1EnAc2D80Zj/wpep5GLguyQ1TnuvqWC8/cP/Y/kbrdd6DVus5XEnk1loY19p81oDNHcZsBc4PPF4C3tVhzFbg+cFBSQ7SewUP8D9Jzl7WbF9vC/BC/8DdvmItn7hv9pXHG13++lzJfN7sWo37mpWO3fVrxnttfa7k62fxvKdx7O7fZ/T5cyU/h2n/HprmfK78OYw+f1bHj6+0o0vQRz3juoIxVNVR4GiH7zl5UslCVc1P41gtcn3Gc33Gc33GW6vr0+WSyxKwfeDxNuDCFYyRJF1FXYJ+CtiVZGeSa4ADwImhMSeAO/r/2uXdwHeq6vnhA0mSrp6Jl1yq6lKSQ8CDwCbgWFWdSXJXf/8R4CRwG7AIfBe48+pN+XumcummYa7PeK7PeK7PeGtyfVL1hkvdkqR1yHeKSlIjDLokNWLdBX3SbQg2oiTPJnk8ySNJFvrbfiTJ3yb51/6vb5/1PFdTkmNJLiZ5YmDbimuS5Pf659TZJB+YzaxXzwrr84kk3+qfR48kuW1g34ZZnyTbk/xdkqeSnEnysf72tX/+VNW6+aD3l7L/BvwEcA3wKLB71vOa9QfwLLBlaNsfAYf7nx8GPjXrea7ymrwXuAV4YtKa0LulxaPAW4Gd/XNs06yfwwzW5xPA744Yu6HWB7gBuKX/+bXA0/01WPPnz3p7hd7lNgTq2Q98sf/5F4Ffnd1UVl9VPQS8OLR5pTXZDxyvqv+rqmfo/WutPasxz1lZYX1WsqHWp6qer6p/6X/+38BT9N75vubPn/UW9JVuMbDRFfA3SU73b68AcH313wvQ//UdM5vd2rHSmnheveZQ/46pxwYuKWzY9UmyA/g54J9YB+fPegt6p1sMbEDvqapb6N318u4k7531hNYZz6ue+4CfBG6mdx+mT/e3b8j1SfI24KvA71TVf40bOmLbTNZnvQXdWwyMUFUX+r9eBL5G74973371jpf9Xy/OboZrxkpr4nkFVNW3q+rlqnoF+FNeu2yw4dYnyVvoxfzPq+r+/uY1f/6st6B3uQ3BhpLkB5Nc++rnwPuBJ+ity0f6wz4CfH02M1xTVlqTE8CBJG9NspPeff3/eQbzm6mhW15/kN55BBtsfZIE+DPgqar644Fda/786XK3xTWjVrgNwYynNWvXA1/rnYNsBv6iqh5Icgr4SpKPAs8Bvz7DOa66JF8GbgW2JFkCPg78ISPWpHq3svgK8CRwCbi7ql6eycRXyQrrc2uSm+ldLngW+C3YkOvzHuDDwONJHulv+33WwfnjW/8lqRHr7ZKLJGkFBl2SGmHQJakRBl2SGmHQJakRBl2SGmHQJakR/w/0mzSs4In2fAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar(X_indices - .25, svm_weights, width=.2, label='SVM weight', color='r')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "feat_imp = pd.DataFrame({\"Col\": x_resampled_encoding.columns})\n",
    "feat_imp[\"Weights\"] = svm_weights\n",
    "\n",
    "#Group OneHotEncoded features\n",
    "feat_imp[\"Col\"] = feat_imp[\"Col\"].apply(lambda x: (x.split('cat')[0]+'cat') if 'cat' in x else x)\n",
    "feat_imp = feat_imp.groupby(\"Col\",as_index=False).sum().sort_values(\"Weights\", ascending=False)\n",
    "#Scale Weights\n",
    "feat_imp[\"FeatImp\"] = (\n",
    "        (feat_imp[\"Weights\"] - feat_imp[\"Weights\"].min())\n",
    "        / (feat_imp[\"Weights\"].max() - feat_imp[\"Weights\"].min())\n",
    "    ).round(4)\n",
    "\n",
    "num_features = 15 \n",
    "plt.figure(figsize=(10,10))\n",
    "plt.title('Feature Importances')\n",
    "# only plot the customized number of features\n",
    "plt.barh(range(num_features), feat_imp['FeatImp'][:num_features][::-1], color='b', align='center')\n",
    "plt.yticks(range(num_features), feat_imp['Col'][:num_features][::-1])\n",
    "plt.xlabel('Relative Importance')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Selected features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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": [
    "clf_selected = svm.SVC(kernel='linear')\n",
    "clf_selected.fit(selector.transform(X), Y)\n",
    "\n",
    "svm_weights_selected = (clf_selected.coef_ ** 2).sum(axis=0)\n",
    "svm_weights_selected /= svm_weights_selected.max()\n",
    "\n",
    "plt.bar(X_indices[selector.get_support()] - .05, svm_weights_selected,\n",
    "        width=.2, label='SVM weights after selection', color='b')\n",
    "\n",
    "\n",
    "plt.title(\"Comparing feature selection\")\n",
    "plt.xlabel('Feature number')\n",
    "plt.yticks(())\n",
    "plt.axis('tight')\n",
    "plt.legend(loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False, False,  True, False,  True,  True, False, False, False,\n",
       "       False, False, False,  True, False,  True, False, False, False,\n",
       "       False, False, False, False, False, False,  True,  True, False,\n",
       "       False, False,  True, False, False,  True, False, False,  True,\n",
       "        True,  True,  True,  True,  True, False, False, False, False,\n",
       "        True,  True, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False,  True,  True, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False,  True,\n",
       "        True, False,  True, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False, False, False, False, False,\n",
       "       False, False, False, False, False])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "selector.get_support()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.86080644, 0.78659234, 1.        , 0.28121704, 0.71196529,\n",
       "       0.01683123, 0.01652907, 0.01197477, 0.016753  , 0.56015671,\n",
       "       0.11837172, 0.13378936, 0.28769452, 0.23225764, 0.46422308,\n",
       "       0.07579656, 0.07579656, 0.00443078, 0.00443078, 0.01528849,\n",
       "       0.01528849, 0.37521584])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_weights_selected "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "feat_imp = pd.DataFrame({\"Col\": x_resampled_encoding.columns})\n",
    "feat_imp[\"SelectedWeights\"]=selector.get_support()\n",
    "feat_selected=feat_imp[feat_imp[\"SelectedWeights\"]== True]\n",
    "feat_selected[\"Weights\"]= svm_weights_selected\n",
    "feat_selected = feat_selected.sort_values(\"Weights\", ascending=False)\n",
    "feat_selected[\"FeatImp\"] = (\n",
    "        (feat_selected[\"Weights\"] - feat_selected[\"Weights\"].min())\n",
    "        / (feat_selected[\"Weights\"].max() -feat_selected[\"Weights\"].min())\n",
    "    ).round(4)\n",
    "\n",
    "num_features = 15 \n",
    "plt.figure(figsize=(10,10))\n",
    "plt.title('Feature Importances')\n",
    "# only plot the customized number of features\n",
    "plt.barh(range(num_features), feat_selected['FeatImp'][:num_features][::-1], color='b', align='center')\n",
    "plt.yticks(range(num_features), feat_selected['Col'][:num_features][::-1])\n",
    "plt.xlabel('Relative Importance')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data with Resampling and without OneHotEncoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time taken to fit the model in sec is  387.84402040000714\n",
      "Accuracy score: 0.8889\n",
      "F1 score : 0.8845\n",
      "Recall score : 0.8484\n",
      "Precision score : 0.9238\n"
     ]
    }
   ],
   "source": [
    "X = x_resampled.values\n",
    "Y = np.squeeze(y_resampled.values)\n",
    "X_train,X_test2,y_train,y_test2=train_test_split(X,Y,test_size=0.2,random_state=10)\n",
    "\n",
    "scaler = StandardScaler().fit(X_train)\n",
    "train_sc = scaler.transform(X_train)\n",
    "test_sc = scaler.transform(X_test2)\n",
    "\n",
    "svm_clf = LinearSVC(random_state=48) \n",
    "start_time = timeit.default_timer()\n",
    "svm_clf.fit(train_sc,y_train)\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print(\"Time taken to fit the model in sec is \",elapsed)\n",
    "y_pred2=svm_clf.predict(test_sc)\n",
    "print(f\"Accuracy score: {accuracy_score(y_test2, y_pred2):.4f}\")\n",
    "print(f\"F1 score : {f1_score(y_test2, y_pred2):.4f}\")\n",
    "print(f\"Recall score : {recall_score(y_test2, y_pred2):.4f}\")\n",
    "print(f\"Precision score : {precision_score(y_test2, y_pred2):.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(confusion_matrix(y_test2,y_pred2),annot=True,fmt='', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.86      0.93      0.89    114386\n",
      "           1       0.92      0.85      0.88    115022\n",
      "\n",
      "    accuracy                           0.89    229408\n",
      "   macro avg       0.89      0.89      0.89    229408\n",
      "weighted avg       0.89      0.89      0.89    229408\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test2, y_pred2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.roc_curve.RocCurveDisplay at 0x2eb1e81cdf0>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "RocCurveDisplay.from_predictions(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.roc_curve.RocCurveDisplay at 0x2eb1e310df0>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "RocCurveDisplay.from_estimator(svm_clf, test_sc, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OneHotEncoding improved accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Change dual and check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time taken to fit the model in sec is  64.24129019999998\n"
     ]
    }
   ],
   "source": [
    "svm_clf = LinearSVC(random_state=48,dual=False) \n",
    "start_time = timeit.default_timer()\n",
    "svm_clf.fit(train_sc,y_train)\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print(\"Time taken to fit the model in sec is \",elapsed)\n",
    "y_pred2=svm_clf.predict(test_sc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy score: 0.9018\n",
      "F1 score : 0.8984\n",
      "Recall score : 0.8664\n",
      "Precision score : 0.9330\n"
     ]
    }
   ],
   "source": [
    "print(f\"Accuracy score: {accuracy_score(y_test2, y_pred2):.4f}\")\n",
    "print(f\"F1 score : {f1_score(y_test2, y_pred2):.4f}\")\n",
    "print(f\"Recall score : {recall_score(y_test2, y_pred2):.4f}\")\n",
    "print(f\"Precision score : {precision_score(y_test2, y_pred2):.4f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Changing dual and maxiter has no impact on accuracy score\n",
    "* Dual=False can resolve the convergence warning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Performing Cross Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time taken to fit the model in sec is  65.19228939999994\n",
      "Time taken to perform cv for the model in sec is  596.9197777000002\n",
      "Cv2 scores: [0.05488911 0.6682243  0.67090947 0.66442321 0.66982843 0.67007253\n",
      " 0.67017428 0.67027889 0.67383591 0.66982555]\n",
      "Mean 5-Fold R Squared: 0.6082461676167825\n"
     ]
    }
   ],
   "source": [
    "svm_clf = LinearSVC(random_state=48,dual=False) \n",
    "start_time = timeit.default_timer()\n",
    "svm_clf.fit(train_sc,y_train)\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print(\"Time taken to fit the model in sec is \",elapsed)\n",
    "\n",
    "start_time = timeit.default_timer()\n",
    "#perform k fold\n",
    "cv_r2_scores_rf = cross_val_score(svm_clf, X, Y, cv=10,scoring='r2')\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print(\"Time taken to perform cv for the model in sec is \",elapsed)\n",
    "\n",
    "print(\"Cv2 scores:\",cv_r2_scores_rf)\n",
    "print(\"Mean 5-Fold R Squared: {}\".format(np.mean(cv_r2_scores_rf)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time taken to fit the model in sec is  66.21286610000243\n",
      "Time taken to perform cv for the model in sec is  257.3644861000066\n",
      "Cv2 scores: [0.36410239 0.66691513 0.66921672 0.66935621 0.67082086]\n",
      "Mean 5-Fold R Squared: 0.6080822606911251\n"
     ]
    }
   ],
   "source": [
    "svm_clf = LinearSVC(random_state=48,dual=False) \n",
    "start_time = timeit.default_timer()\n",
    "svm_clf.fit(train_sc,y_train)\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print(\"Time taken to fit the model in sec is \",elapsed)\n",
    "\n",
    "start_time = timeit.default_timer()\n",
    "#perform k fold\n",
    "cv_r2_scores_rf = cross_val_score(svm_clf, X, Y, cv=5,scoring='r2')\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print(\"Time taken to perform cv for the model in sec is \",elapsed)\n",
    "\n",
    "print(\"Cv2 scores:\",cv_r2_scores_rf)\n",
    "print(\"Mean 5-Fold R Squared: {}\".format(np.mean(cv_r2_scores_rf)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* We see that the metrics for all folds in cross_val_score are uniform so it can be concluded that the model is able to generalise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "provenance": []
  },
  "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}