Pipe_Line_Logistic_Regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "366b4301",
   "metadata": {},
   "source": [
    "# بسم الله الرحمن الرحيم"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "2144698b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils import *\n",
    "import pandas as pd \n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import accuracy_score\n",
    "from matplotlib import pyplot as plt \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "3acb315d",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"dataset/body_level_classification_train.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "09063b7a",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_original= cleaning_data(df).copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "8323be7f",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_class_train,all_class_test=data_spliting(df_original,1/3)\n",
    "## shuffling \n",
    "all_class_train = all_class_train.sample(frac=1.0, random_state=42)\n",
    "all_class_test = all_class_test.sample(frac=1.0, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "4742b3ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train=all_class_train.drop([\"Body_Level\"],axis=1).to_numpy()\n",
    "Y_train=all_class_train[\"Body_Level\"].to_numpy()\n",
    "X_test =all_class_test.drop([\"Body_Level\"],axis=1).to_numpy()\n",
    "Y_test =all_class_test[\"Body_Level\"].to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "47964cd4",
   "metadata": {},
   "outputs": [],
   "source": [
    "class_weights = all_class_train[\"Body_Level\"].value_counts().to_dict()\n",
    "total_num = sum(class_weights.values())\n",
    "class_weights= { i : (total_num/(4* j )) for i ,j in zip(class_weights.keys(),class_weights.values())}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "a42bc0bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, scaler= data_scaling(X_train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "65948699",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_val, X_test, Y_val, Y_test = train_test_split(X_test\n",
    "                                                     , Y_test, test_size=0.5, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "099adff6",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_val = scaler.transform(X_val)\n",
    "X_test=scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d6e5e21",
   "metadata": {},
   "source": [
    "# Bare Logistic Regression "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "ce925de0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression()"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bare_reg_model = LogisticRegression()\n",
    "# concat the train and the val becuase it's the bare one \n",
    "bare_reg_model.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "382f7596",
   "metadata": {},
   "source": [
    "## accuracy on test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "6f33ccee",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9533468559837728\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.91      1.00      0.95       127\n",
      "           1       0.94      0.74      0.83       134\n",
      "           2       0.91      0.97      0.94       271\n",
      "           3       1.00      0.99      0.99       454\n",
      "\n",
      "    accuracy                           0.95       986\n",
      "   macro avg       0.94      0.93      0.93       986\n",
      "weighted avg       0.95      0.95      0.95       986\n",
      "\n"
     ]
    }
   ],
   "source": [
    "bare_accuracy , repo = test_model(bare_reg_model,X_train,Y_train,axis=False)\n",
    "print(bare_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "2d114096",
   "metadata": {},
   "outputs": [],
   "source": [
    "bare_accuracy , repo = test_model(bare_reg_model,X_test,Y_test,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "7c95ae46",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9512195121951219\n"
     ]
    }
   ],
   "source": [
    "print(bare_accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "20a5e8c0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.96      0.92        28\n",
      "           1       0.96      0.72      0.83        36\n",
      "           2       0.93      0.99      0.96        80\n",
      "           3       0.99      1.00      1.00       102\n",
      "\n",
      "    accuracy                           0.95       246\n",
      "   macro avg       0.94      0.92      0.92       246\n",
      "weighted avg       0.95      0.95      0.95       246\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(repo)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9abd1d4b",
   "metadata": {},
   "source": [
    "## Regression model with class weights only \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "58b05a31",
   "metadata": {},
   "outputs": [],
   "source": [
    "only_weights_model = LogisticRegression(random_state=42,max_iter=3000\n",
    "                                        ,class_weight=class_weights)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "be8aaf95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(class_weight={0: 1.9409448818897639, 1: 1.8395522388059702,\n",
       "                                 2: 0.9095940959409594, 3: 0.5429515418502202},\n",
       "                   max_iter=3000, random_state=42)"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "only_weights_model.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "34724413",
   "metadata": {},
   "outputs": [],
   "source": [
    "only_weights_accuracy , repo = test_model(only_weights_model,X_test,Y_test,axis=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "6cd50756",
   "metadata": {},
   "outputs": [],
   "source": [
    "only_weights_accuracy , repo= test_model(only_weights_model,X_train,Y_train,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "7fca9383",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.960446247464503\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      1.00      0.96       127\n",
      "           1       0.91      0.87      0.89       134\n",
      "           2       0.95      0.95      0.95       271\n",
      "           3       1.00      0.98      0.99       454\n",
      "\n",
      "    accuracy                           0.96       986\n",
      "   macro avg       0.94      0.95      0.95       986\n",
      "weighted avg       0.96      0.96      0.96       986\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(only_weights_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84b5fbe2",
   "metadata": {},
   "source": [
    "## applying the hyperpamater tuning \n",
    "**In this model we get the best we can get from the Logistic regression by applying the Tuning and get the best hyperparameters**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "293e51a9",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 2 folds for each of 12 candidates, totalling 24 fits\n",
      "[CV] END ................................C=0.001, penalty=l2; total time=   0.0s\n",
      "[CV] END ................................C=0.001, penalty=l2; total time=   0.0s\n",
      "[CV] END .................................C=0.01, penalty=l2; total time=   0.0s\n",
      "[CV] END .................................C=0.01, penalty=l2; total time=   0.0s\n",
      "[CV] END ..................................C=0.1, penalty=l2; total time=   0.0s\n",
      "[CV] END ..................................C=0.1, penalty=l2; total time=   0.0s\n",
      "[CV] END ....................................C=1, penalty=l2; total time=   0.0s\n",
      "[CV] END ....................................C=1, penalty=l2; total time=   0.0s\n",
      "[CV] END ...................................C=10, penalty=l2; total time=   0.1s\n",
      "[CV] END ...................................C=10, penalty=l2; total time=   0.1s\n",
      "[CV] END ..................................C=100, penalty=l2; total time=   0.1s\n",
      "[CV] END ..................................C=100, penalty=l2; total time=   0.1s\n",
      "[CV] END ..................................C=200, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=200, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=300, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=300, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=400, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=400, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=500, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=500, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=600, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=600, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=700, penalty=l2; total time=   0.2s\n",
      "[CV] END ..................................C=700, penalty=l2; total time=   0.2s\n",
      "0.001 0.49878576707845\n",
      "0.01 0.6702077394150565\n",
      "0.1 0.8334798331749551\n",
      "1 0.8984558124802027\n",
      "10 0.946377098511245\n",
      "100 0.9642500791891035\n",
      "200 0.9650630873191849\n",
      "300 0.9675007918910358\n",
      "400 0.9675007918910358\n",
      "500 0.9683138000211171\n",
      "600 0.9691254883328054\n",
      "700 0.9691268081511983\n"
     ]
    }
   ],
   "source": [
    "\n",
    "reg_model = LogisticRegression(random_state=42,max_iter=3000)\n",
    "# Define the parameter grid to search over\n",
    "param_grid = {\n",
    "    'C': [0.001,0.01,0.1, 1, 10,100,200,300,400,500 ,600 ,700 ],\n",
    "    'penalty': ['l2']\n",
    "}\n",
    "\n",
    "\n",
    "grid_search = GridSearchCV(estimator=reg_model, param_grid=param_grid,\n",
    "                           cv=2,verbose=2)\n",
    "\n",
    "# concat beacuase we use the K-Fold cross validation \n",
    "grid_search.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))\n",
    "\n",
    "\n",
    "\n",
    "# Access the cv_results_ dictionary\n",
    "cv_results = grid_search.cv_results_\n",
    "\n",
    "# Print the mean test scores for each hyperparameter combination\n",
    "for mean_score, params in zip(cv_results['mean_test_score'], cv_results['params']):\n",
    "    print(params['C'], mean_score)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "11d8b4b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_reg_model = grid_search.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "71d13a65",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "y_train_pred = best_reg_model.predict(X_train)\n",
    "training_error = mean_squared_error(Y_train, y_train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "070f70bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_results = grid_search.cv_results_\n",
    "validation_scores = cv_results['mean_test_score']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "id": "4dc352ac",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Error: 0.0010141987829614604\n",
      "Validation Score (Fold 1): 0.49878576707845\n",
      "Validation Score (Fold 2): 0.6702077394150565\n",
      "Validation Score (Fold 3): 0.8334798331749551\n",
      "Validation Score (Fold 4): 0.8984558124802027\n",
      "Validation Score (Fold 5): 0.946377098511245\n",
      "Validation Score (Fold 6): 0.9642500791891035\n",
      "Validation Score (Fold 7): 0.9650630873191849\n",
      "Validation Score (Fold 8): 0.9675007918910358\n",
      "Validation Score (Fold 9): 0.9675007918910358\n",
      "Validation Score (Fold 10): 0.9683138000211171\n",
      "Validation Score (Fold 11): 0.9691254883328054\n",
      "Validation Score (Fold 12): 0.9691268081511983\n"
     ]
    }
   ],
   "source": [
    "val_error=[]\n",
    "print(\"Training Error:\", training_error)\n",
    "for fold, score in enumerate(validation_scores):\n",
    "    print(\"Validation Score (Fold {}):\".format(fold+1), score)\n",
    "    val_error.append(1-score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "id": "a239ddd5",
   "metadata": {},
   "outputs": [],
   "source": [
    "C_values = [0.001,0.01,0.1, 1, 10,100,200,300,400,500 ,600 ,700]\n",
    "train_error = []\n",
    "\n",
    "for c in C_values: \n",
    "    sample_reg=LogisticRegression(C=c,max_iter=2000)\n",
    "    sample_reg.fit(X_train,Y_train)\n",
    "    y_train_pred = sample_reg.predict(X_train)\n",
    "    training_error = mean_squared_error(Y_train, y_train_pred)\n",
    "    train_error.append(training_error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "491591a2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "2377b68e",
   "metadata": {},
   "source": [
    "## accuracy on test on best reg model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "id": "7a8a240d",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_accuracy , repo = test_model(best_reg_model,X_test,Y_test,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "id": "f8da8533",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.983739837398374\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      0.96      0.96        28\n",
      "           1       0.97      0.97      0.97        36\n",
      "           2       1.00      0.97      0.99        80\n",
      "           3       0.98      1.00      0.99       102\n",
      "\n",
      "    accuracy                           0.98       246\n",
      "   macro avg       0.98      0.98      0.98       246\n",
      "weighted avg       0.98      0.98      0.98       246\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(best_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "001bc369",
   "metadata": {},
   "source": [
    "## Plotting the change on accuracy tuning the 'C' hyperparameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "id": "f04980ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Print the mean test scores for each hyperparameter combination\n",
    "x_axis = []\n",
    "y_axis = []\n",
    "for mean_score, params in zip(cv_results['mean_test_score'], cv_results['params']):\n",
    "    x_axis.append(params[\"C\"])\n",
    "    y_axis.append(mean_score)\n",
    "    \n",
    "    \n",
    "plt.plot(np.arange(1,len(x_axis)+1),y_axis)\n",
    "plt.title('Changing the hyperparameter C vs accuracy')\n",
    "plt.xlabel('C')\n",
    "plt.ylabel('val accuracy')\n",
    "plt.axvline(np.argmax(cv_results[\"mean_test_score\"]),\n",
    "            color='red',\n",
    "            label=f\"C={cv_results['params'][np.argmax(cv_results['mean_test_score'])]['C']}\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "id": "54158722",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(np.arange(1,len(x_axis)+1),val_error,label='val_error')\n",
    "plt.plot(np.arange(1,len(x_axis)+1),train_error,label='train_error',color='red')\n",
    "plt.title(\"train vs val error while changing the value of C \")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69be64ab",
   "metadata": {},
   "source": [
    "## Regression model with best C only\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "id": "42ab0772",
   "metadata": {},
   "outputs": [],
   "source": [
    "only_C_model = LogisticRegression(random_state=42,max_iter=3000\n",
    "                                        ,C=700)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "id": "45d874d9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=700, max_iter=3000, random_state=42)"
      ]
     },
     "execution_count": 188,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "only_C_model.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "id": "c04e839b",
   "metadata": {},
   "outputs": [],
   "source": [
    "only_C_accuracy , repo = test_model(only_C_model,X_test,Y_test,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "id": "e9b21110",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.983739837398374\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      0.96      0.96        28\n",
      "           1       0.97      0.97      0.97        36\n",
      "           2       1.00      0.97      0.99        80\n",
      "           3       0.98      1.00      0.99       102\n",
      "\n",
      "    accuracy                           0.98       246\n",
      "   macro avg       0.98      0.98      0.98       246\n",
      "weighted avg       0.98      0.98      0.98       246\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(only_C_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "092b4370",
   "metadata": {},
   "source": [
    "**As shown from above analysis , it's obvious that the Tuning the Parameter C alone , or providing class_weights alone is much better that apply both of them**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d748cfa4",
   "metadata": {},
   "source": [
    "# oversampling "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "id": "cc58b918",
   "metadata": {},
   "outputs": [],
   "source": [
    "oversampled= over_sampling(all_class_train,sampling_ratio=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "e3c7e779",
   "metadata": {},
   "outputs": [],
   "source": [
    "class_weights = oversampled[\"Body_Level\"].value_counts().to_dict()\n",
    "total_num = sum(class_weights.values())\n",
    "class_weights= { i : (total_num/(4* j )) for i ,j in zip(class_weights.keys(),class_weights.values())}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "id": "6ff5a061",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{3: 0.7252202643171806,\n",
       " 2: 0.9571220930232558,\n",
       " 1: 1.256679389312977,\n",
       " 0: 1.2811284046692606}"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "id": "8dc52235",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train=oversampled.drop([\"Body_Level\"],axis=1).to_numpy()\n",
    "Y_train=oversampled[\"Body_Level\"].to_numpy()\n",
    "X_test =all_class_test.drop([\"Body_Level\"],axis=1).to_numpy()\n",
    "Y_test =all_class_test[\"Body_Level\"].to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "id": "938a5761",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, scaler= data_scaling(X_train)\n",
    "X_val, X_test, Y_val, Y_test = train_test_split(X_test\n",
    "                                                     , Y_test, test_size=0.5, random_state=42)\n",
    "X_val = scaler.transform(X_val)\n",
    "X_test=scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5a22c9ec",
   "metadata": {},
   "source": [
    "## Applying the bare Regression on Oversampled data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "id": "e876c131",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(max_iter=1000)"
      ]
     },
     "execution_count": 196,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bare_oversampled_reg = LogisticRegression(max_iter=1000)\n",
    "bare_oversampled_reg.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "id": "d518d798",
   "metadata": {},
   "outputs": [],
   "source": [
    "bare_oversampled_accuracy , repo = test_model(bare_oversampled_reg,X_test,Y_test,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "dd678bab",
   "metadata": {},
   "outputs": [],
   "source": [
    "bare_oversampled_accuracy , repo = test_model(bare_oversampled_reg,X_train,Y_train,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "id": "37645aa8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9552012148823082\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.91      1.00      0.96       257\n",
      "           1       0.96      0.84      0.90       262\n",
      "           2       0.94      0.96      0.95       344\n",
      "           3       0.99      0.99      0.99       454\n",
      "\n",
      "    accuracy                           0.96      1317\n",
      "   macro avg       0.95      0.95      0.95      1317\n",
      "weighted avg       0.96      0.96      0.95      1317\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(bare_oversampled_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09091c9a",
   "metadata": {},
   "source": [
    "## Applying Regression On Over sampled data Class_weights only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "id": "9affc988",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(class_weight={0: 1.2811284046692606, 1: 1.256679389312977,\n",
       "                                 2: 0.9571220930232558, 3: 0.7252202643171806},\n",
       "                   max_iter=1000)"
      ]
     },
     "execution_count": 200,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oversampled_class_weights = LogisticRegression(max_iter=1000,class_weight=class_weights)\n",
    "oversampled_class_weights.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "id": "8b5c4ad4",
   "metadata": {},
   "outputs": [],
   "source": [
    "oversampled_class_weights_accuracy , repo = test_model(oversampled_class_weights,X_test,\n",
    "                                                       Y_test,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "id": "ec5246ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "oversampled_class_weights_accuracy , repo = test_model(oversampled_class_weights,X_train,\n",
    "                                                       Y_train,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "id": "0b01b544",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9597570235383447\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      1.00      0.96       257\n",
      "           1       0.95      0.88      0.91       262\n",
      "           2       0.96      0.95      0.95       344\n",
      "           3       0.99      0.99      0.99       454\n",
      "\n",
      "    accuracy                           0.96      1317\n",
      "   macro avg       0.95      0.96      0.95      1317\n",
      "weighted avg       0.96      0.96      0.96      1317\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(oversampled_class_weights_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7e3ffcfb",
   "metadata": {},
   "source": [
    "## Tuning the C value on oversampled data  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "id": "368eac1b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 15 candidates, totalling 75 fits\n",
      "{'C': 0.001, 'penalty': 'l2'} 0.5659457688211681\n",
      "{'C': 0.01, 'penalty': 'l2'} 0.7182948308347669\n",
      "{'C': 0.1, 'penalty': 'l2'} 0.8655505038092898\n",
      "{'C': 1, 'penalty': 'l2'} 0.9391701482755795\n",
      "{'C': 10, 'penalty': 'l2'} 0.9705353485704924\n",
      "{'C': 100, 'penalty': 'l2'} 0.9807815188006881\n",
      "{'C': 200, 'penalty': 'l2'} 0.9820635700827394\n",
      "{'C': 300, 'penalty': 'l2'} 0.9833435733595477\n",
      "{'C': 400, 'penalty': 'l2'} 0.9839845990005734\n",
      "{'C': 500, 'penalty': 'l2'} 0.9839845990005734\n",
      "{'C': 600, 'penalty': 'l2'} 0.984625624641599\n",
      "{'C': 700, 'penalty': 'l2'} 0.984625624641599\n",
      "{'C': 800, 'penalty': 'l2'} 0.984625624641599\n",
      "{'C': 900, 'penalty': 'l2'} 0.9839866470058164\n",
      "{'C': 1000, 'penalty': 'l2'} 0.9839866470058164\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "\n",
    "\n",
    "reg_model = LogisticRegression(random_state=42,max_iter=3000)\n",
    "# Define the parameter grid to search over\n",
    "param_grid = {\n",
    "    'C': [0.001,0.01,0.1, 1, 10,100,200,300,400,500 ,600 ,700,800,900,1000],\n",
    "    'penalty': ['l2']\n",
    "}\n",
    "\n",
    "\n",
    "grid_search = GridSearchCV(estimator=reg_model, param_grid=param_grid, cv=5 , verbose=1 )\n",
    "\n",
    "grid_search.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))\n",
    "\n",
    "\n",
    "\n",
    "# Access the cv_results_ dictionary\n",
    "cv_results = grid_search.cv_results_\n",
    "\n",
    "# Print the mean test scores for each hyperparameter combination\n",
    "for mean_score, params in zip(cv_results['mean_test_score'], cv_results['params']):\n",
    "    print(params, mean_score)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "id": "00809166",
   "metadata": {},
   "outputs": [],
   "source": [
    "Oversampled_C_only_accuracy , repo = test_model(grid_search.best_estimator_,\n",
    "                                                X_test,Y_test,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "id": "8f36b13b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.983739837398374\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      0.96      0.96        28\n",
      "           1       0.97      0.97      0.97        36\n",
      "           2       1.00      0.97      0.99        80\n",
      "           3       0.98      1.00      0.99       102\n",
      "\n",
      "    accuracy                           0.98       246\n",
      "   macro avg       0.98      0.98      0.98       246\n",
      "weighted avg       0.98      0.98      0.98       246\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(Oversampled_C_only_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92c4f0df",
   "metadata": {},
   "source": [
    "## Plot the different values of C "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "id": "ac9df260",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_axis = []\n",
    "y_axis = []\n",
    "for mean_score, params in zip(cv_results['mean_test_score'], cv_results['params']):\n",
    "    x_axis.append(params[\"C\"])\n",
    "    y_axis.append(mean_score)\n",
    "    \n",
    "    \n",
    "plt.plot(np.arange(1,len(x_axis)+1),y_axis)\n",
    "plt.title('Changing the hyperparameter C vs accuracy')\n",
    "plt.xlabel('C')\n",
    "plt.ylabel('accuracy')\n",
    "plt.axvline(np.argmax(cv_results[\"mean_test_score\"]),\n",
    "            color='red',\n",
    "            label=f\"C={cv_results['params'][np.argmax(cv_results['mean_test_score'])]['C']}\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc50d56a",
   "metadata": {},
   "source": [
    "## Logistic Regression on best C and class weights on Oversampled data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "9e7ab03f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=600,\n",
       "                   class_weight={0: 1.2811284046692606, 1: 1.256679389312977,\n",
       "                                 2: 0.9571220930232558, 3: 0.7252202643171806},\n",
       "                   max_iter=1000)"
      ]
     },
     "execution_count": 208,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_oversampled_class_weights = LogisticRegression(max_iter=1000,class_weight=class_weights,C=600)\n",
    "best_oversampled_class_weights.fit(np.concatenate([X_train,X_val], axis=0)\n",
    "                   ,np.concatenate([Y_train,Y_val],axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "id": "2bddf26a",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_oversampled_class_weights_accuracy , repo = test_model(best_oversampled_class_weights,\n",
    "                                                            X_test,Y_test,axis=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "e1169e3b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9878048780487805\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.97      1.00      0.98        28\n",
      "           1       1.00      0.97      0.99        36\n",
      "           2       1.00      0.97      0.99        80\n",
      "           3       0.98      1.00      0.99       102\n",
      "\n",
      "    accuracy                           0.99       246\n",
      "   macro avg       0.99      0.99      0.99       246\n",
      "weighted avg       0.99      0.99      0.99       246\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(best_oversampled_class_weights_accuracy)\n",
    "print(repo)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "id": "57f25b74",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"Best_Logistic_Regression_Model.h5\",\"wb\") as fb: \n",
    "            pickle.dump(best_oversampled_class_weights,fb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "id": "b3c9d836",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar([\"bare_reg\",\"only CW\",\"best C\",\"OS\",\n",
    "        \"OS+ CW\",\"OS+best C \",\"OS+best C + CW\"],\n",
    "        [bare_accuracy,only_weights_accuracy,only_C_accuracy,\n",
    "        bare_oversampled_accuracy,oversampled_class_weights_accuracy,Oversampled_C_only_accuracy \n",
    "        ,best_oversampled_class_weights_accuracy ]\n",
    "        ,width=0.25,)\n",
    "plt.title(\"title='different approches VS accuracy\")\n",
    "plt.xticks(rotation='vertical')\n",
    "plt.axhline(bare_accuracy, color='red',label='bare reg acc') \n",
    "plt.legend(loc=[1,1.1])\n",
    "plt.xlabel('Different Approch ')\n",
    "plt.ylabel('Accuracy')\n",
    "\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d00bfa54",
   "metadata": {},
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}