{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "985e5184",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score, roc_auc_score, classification_report\n",
    "import statsmodels.api as sm\n",
    "import matplotlib.pyplot as plt\n",
    "from statsmodels.stats.outliers_influence import variance_inflation_factor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e5ce7868",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load Flier Response Data\n",
    "flierresponse = pd.read_csv(\"D:\\Academics\\BTech\\DSC4.51015_StatisticalModeling-1\\Session 06\\FlierResponse.csv\")\n",
    "\n",
    "# Convert Response column to factor\n",
    "flierresponse['Response'] = pd.factorize(flierresponse['Response'])[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "99f9efd9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              Results: Generalized linear model\n",
      "==============================================================\n",
      "Model:              GLM              AIC:            53.9366  \n",
      "Link Function:      Logit            BIC:            -357.0243\n",
      "Dependent Variable: Response         Log-Likelihood: -24.968  \n",
      "Date:               2024-03-26 14:58 LL-Null:        -61.578  \n",
      "No. Observations:   92               Deviance:       49.937   \n",
      "Df Model:           1                Pearson chi2:   46.3     \n",
      "Df Residuals:       90               Scale:          1.0000   \n",
      "Method:             IRLS                                      \n",
      "---------------------------------------------------------------\n",
      "            Coef.   Std.Err.     z     P>|z|    [0.025   0.975]\n",
      "---------------------------------------------------------------\n",
      "const      20.4078    4.5233   4.5117  0.0000  11.5423  29.2734\n",
      "Age        -0.4259    0.0948  -4.4921  0.0000  -0.6118  -0.2401\n",
      "==============================================================\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Logistic Regression Model\n",
    "X = flierresponse[['Age']]\n",
    "y = flierresponse['Response']\n",
    "flierresponseglm = sm.GLM(y, sm.add_constant(X), family=sm.families.Binomial()).fit()\n",
    "print(flierresponseglm.summary2())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "78f3cffb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of nd_array: (1, 1)\n",
      "Shape of flierresponseglm.params: (2,)\n"
     ]
    }
   ],
   "source": [
    "# Predict probability for Age=50\n",
    "nd = pd.DataFrame({'Age': [50]})\n",
    "print(\"Shape of nd_array:\", nd_array.shape)\n",
    "print(\"Shape of flierresponseglm.params:\", flierresponseglm.params.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "ad4b20d5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.29152877]\n"
     ]
    }
   ],
   "source": [
    "# Construct the design matrix with a constant term manually\n",
    "nd_array = np.column_stack((np.ones_like(nd['Age']), nd['Age']))\n",
    "\n",
    "# Predict the response for the new data\n",
    "print(flierresponseglm.predict(nd_array))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "65942c08",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.29152877]\n"
     ]
    }
   ],
   "source": [
    "# Construct the design matrix with a constant term manually\n",
    "nd_array = np.column_stack((np.ones_like(nd['Age']), nd['Age']))\n",
    "\n",
    "# Predict the response for the new data with transform=False\n",
    "print(flierresponseglm.predict(nd_array, transform=False))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "cb502efc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                 Generalized Linear Model Regression Results                  \n",
      "==============================================================================\n",
      "Dep. Variable:               Response   No. Observations:                   92\n",
      "Model:                            GLM   Df Residuals:                       90\n",
      "Model Family:                Binomial   Df Model:                            1\n",
      "Link Function:                  Logit   Scale:                          1.0000\n",
      "Method:                          IRLS   Log-Likelihood:                -24.968\n",
      "Date:                Tue, 26 Mar 2024   Deviance:                       49.937\n",
      "Time:                        21:02:50   Pearson chi2:                     46.3\n",
      "No. Iterations:                     7   Pseudo R-squ. (CS):             0.5488\n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "const         20.4078      4.523      4.512      0.000      11.542      29.273\n",
      "Age           -0.4259      0.095     -4.492      0.000      -0.612      -0.240\n",
      "==============================================================================\n"
     ]
    }
   ],
   "source": [
    "print(flierresponseglm.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "345b0a1e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-24.96831445537559\n"
     ]
    }
   ],
   "source": [
    "# Print log-likelihood of the model\n",
    "print(flierresponseglm.llf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "cd96a5a4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "49.93662891075118\n",
      "53.93662891075118\n"
     ]
    }
   ],
   "source": [
    "print(flierresponseglm.deviance)\n",
    "print(flierresponseglm.aic)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "61be6eec",
   "metadata": {},
   "source": [
    "# Case Study: Framingham Heart Study"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "ed3c5925",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Read the Framingham dataset\n",
    "framingham = pd.read_csv(r\"D:\\Academics\\BTech\\DSC4.51015_StatisticalModeling-1\\Session 06\\framingham.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "2f2afc94",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 3658 entries, 0 to 4239\n",
      "Data columns (total 16 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   male             3658 non-null   int64  \n",
      " 1   age              3658 non-null   int64  \n",
      " 2   education        3658 non-null   float64\n",
      " 3   currentSmoker    3658 non-null   int64  \n",
      " 4   cigsPerDay       3658 non-null   float64\n",
      " 5   BPMeds           3658 non-null   float64\n",
      " 6   prevalentStroke  3658 non-null   int64  \n",
      " 7   prevalentHyp     3658 non-null   int64  \n",
      " 8   diabetes         3658 non-null   int64  \n",
      " 9   totChol          3658 non-null   float64\n",
      " 10  sysBP            3658 non-null   float64\n",
      " 11  diaBP            3658 non-null   float64\n",
      " 12  BMI              3658 non-null   float64\n",
      " 13  heartRate        3658 non-null   float64\n",
      " 14  glucose          3658 non-null   float64\n",
      " 15  TenYearCHD       3658 non-null   int64  \n",
      "dtypes: float64(9), int64(7)\n",
      "memory usage: 485.8 KB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# Remove rows with missing values from the DataFrame\n",
    "framingham = framingham.dropna()\n",
    "\n",
    "print(framingham.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "d5deec72",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Randomly split the data into training and testing sets\n",
    "train, test = train_test_split(framingham, test_size=0.3, random_state=123)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "5c43757c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Proportion of TenYearCHD in the training dataset: 0.147265625\n",
      "Proportion of TenYearCHD in the testing dataset: 0.16393442622950818\n"
     ]
    }
   ],
   "source": [
    "# Calculate the proportion of TenYearCHD in the training dataset\n",
    "train_proportion = train['TenYearCHD'].sum() / len(train)\n",
    "\n",
    "# Calculate the proportion of TenYearCHD in the testing dataset\n",
    "test_proportion = test['TenYearCHD'].sum() / len(test)\n",
    "\n",
    "print(\"Proportion of TenYearCHD in the training dataset:\", train_proportion)\n",
    "print(\"Proportion of TenYearCHD in the testing dataset:\", test_proportion)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "a8df908b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                Results: Generalized linear model\n",
      "=================================================================\n",
      "Model:               GLM              AIC:            1919.4621  \n",
      "Link Function:       Logit            BIC:            -18077.2457\n",
      "Dependent Variable:  TenYearCHD       Log-Likelihood: -943.73    \n",
      "Date:                2024-03-27 11:35 LL-Null:        -1069.9    \n",
      "No. Observations:    2560             Deviance:       1887.5     \n",
      "Df Model:            15               Pearson chi2:   2.45e+03   \n",
      "Df Residuals:        2544             Scale:          1.0000     \n",
      "Method:              IRLS                                        \n",
      "-----------------------------------------------------------------\n",
      "                 Coef.  Std.Err.    z     P>|z|   [0.025   0.975]\n",
      "-----------------------------------------------------------------\n",
      "const           -8.9157   0.8674 -10.2782 0.0000 -10.6158 -7.2155\n",
      "male             0.5311   0.1324   4.0102 0.0001   0.2715  0.7907\n",
      "age              0.0649   0.0081   8.0364 0.0000   0.0490  0.0807\n",
      "education       -0.0716   0.0604  -1.1857 0.2357  -0.1899  0.0467\n",
      "currentSmoker    0.1411   0.1905   0.7411 0.4586  -0.2321  0.5144\n",
      "cigsPerDay       0.0212   0.0077   2.7441 0.0061   0.0061  0.0363\n",
      "BPMeds           0.2710   0.2872   0.9434 0.3455  -0.2920  0.8339\n",
      "prevalentStroke  0.8431   0.5448   1.5476 0.1217  -0.2246  1.9108\n",
      "prevalentHyp    -0.0460   0.1694  -0.2713 0.7861  -0.3780  0.2861\n",
      "diabetes        -0.0005   0.3976  -0.0013 0.9990  -0.7797  0.7787\n",
      "totChol          0.0034   0.0014   2.4237 0.0154   0.0006  0.0061\n",
      "sysBP            0.0175   0.0047   3.7571 0.0002   0.0084  0.0266\n",
      "diaBP           -0.0040   0.0078  -0.5083 0.6113  -0.0193  0.0114\n",
      "BMI              0.0136   0.0155   0.8793 0.3793  -0.0168  0.0440\n",
      "heartRate       -0.0041   0.0051  -0.8020 0.4226  -0.0142  0.0059\n",
      "glucose          0.0060   0.0026   2.2853 0.0223   0.0009  0.0112\n",
      "=================================================================\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Logistic Regression Model\n",
    "framingham1 = sm.GLM(train['TenYearCHD'], sm.add_constant(train.drop(columns=['TenYearCHD'])),\n",
    "                     family=sm.families.Binomial()).fit()\n",
    "print(framingham1.summary2())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "6d671253",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "VIF:\n",
      "     Variable       VIF\n",
      "0         age  1.335089\n",
      "1   education  1.053133\n",
      "2  cigsPerDay  1.049092\n",
      "3     totChol  1.108764\n",
      "4       sysBP  3.126199\n",
      "5       diaBP  2.875258\n",
      "6         BMI  1.207163\n",
      "7   heartRate  1.065791\n",
      "8     glucose  1.038740\n"
     ]
    }
   ],
   "source": [
    "# Check for multicollinearity\n",
    "\n",
    "X = framingham.drop(columns=['TenYearCHD']+['male']+['currentSmoker']+['BPMeds']+\n",
    "                    ['prevalentStroke']+['prevalentHyp']+['diabetes'], axis=1)\n",
    "# Add a constant term for the intercept\n",
    "X_with_constant = sm.add_constant(X)\n",
    "# Calculate VIF (excluding the constant term)\n",
    "vif_data = pd.DataFrame()\n",
    "vif_data[\"Variable\"] = X.columns\n",
    "vif_data[\"VIF\"] = [variance_inflation_factor(X_with_constant.values, i) \n",
    "                   for i in range(1, X_with_constant.shape[1])]\n",
    "\n",
    "print(\"VIF:\")\n",
    "print(vif_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "6a2e4e4d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix_Train:\n",
      "[[2169   14]\n",
      " [ 346   31]]\n",
      "Accuracy_Train: 0.859375\n",
      "Confusion Matrix_Test:\n",
      "[[913   5]\n",
      " [166  14]]\n",
      "Accuracy_Test: 0.8442622950819673\n"
     ]
    }
   ],
   "source": [
    "# Accuracy on the training set\n",
    "predictTrain1 = framingham1.predict(sm.add_constant(train.drop(columns=['TenYearCHD'])))\n",
    "print(\"Confusion Matrix_Train:\")\n",
    "print(confusion_matrix(train['TenYearCHD'], predictTrain1 > 0.5))\n",
    "print(\"Accuracy_Train:\", accuracy_score(train['TenYearCHD'], predictTrain1 > 0.5))\n",
    "\n",
    "# Accuracy on Test Set\n",
    "predictTest1 = framingham1.predict(sm.add_constant(test.drop(columns=['TenYearCHD'])))\n",
    "print(\"Confusion Matrix_Test:\")\n",
    "print(confusion_matrix(test['TenYearCHD'], predictTest1 > 0.5))\n",
    "print(\"Accuracy_Test:\", accuracy_score(test['TenYearCHD'], predictTest1 > 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "415076d2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "Predicted     0   1\n",
      "Actual             \n",
      "0          2169  14\n",
      "1           346  31\n"
     ]
    }
   ],
   "source": [
    "# Displaying confusion matrix with row and column labels\n",
    "\n",
    "threshold = 0.5  # Example threshold to be adjusted as needed\n",
    "predicted_labels = (predictTrain1 > threshold).astype(int)\n",
    "\n",
    "# Create a DataFrame containing true and predicted labels\n",
    "confusion_data = pd.DataFrame({'Actual': train['TenYearCHD'], 'Predicted': predicted_labels})\n",
    "\n",
    "# Calculate confusion matrix using crosstab\n",
    "conf_matrix = pd.crosstab(confusion_data['Actual'], confusion_data['Predicted'], rownames=['Actual'], colnames=['Predicted'])\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(conf_matrix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "a604670c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix with threshold of 0.1\n",
      "Predicted    0    1\n",
      "Actual             \n",
      "0          442  476\n",
      "1           33  147\n"
     ]
    }
   ],
   "source": [
    "# Confusion matrix on Test Data at various thresholds\n",
    "\n",
    "threshold = 0.1  # Example threshold to be adjusted as needed\n",
    "predicted_labels = (predictTest1 > threshold).astype(int)\n",
    "\n",
    "# Create a DataFrame containing true and predicted labels\n",
    "confusion_data = pd.DataFrame({'Actual': test['TenYearCHD'], 'Predicted': predicted_labels})\n",
    "\n",
    "# Calculate confusion matrix using crosstab\n",
    "conf_matrix = pd.crosstab(confusion_data['Actual'], confusion_data['Predicted'], rownames=['Actual'], colnames=['Predicted'])\n",
    "\n",
    "print(\"Confusion Matrix with threshold of\", threshold)\n",
    "print(conf_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fede107c",
   "metadata": {},
   "source": [
    "# Build another model assuming Education to be a nominal variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "aa167610",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 3658 entries, 0 to 4239\n",
      "Data columns (total 16 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   male             3658 non-null   int64  \n",
      " 1   age              3658 non-null   int64  \n",
      " 2   education        3658 non-null   float64\n",
      " 3   currentSmoker    3658 non-null   int64  \n",
      " 4   cigsPerDay       3658 non-null   float64\n",
      " 5   BPMeds           3658 non-null   float64\n",
      " 6   prevalentStroke  3658 non-null   int64  \n",
      " 7   prevalentHyp     3658 non-null   int64  \n",
      " 8   diabetes         3658 non-null   int64  \n",
      " 9   totChol          3658 non-null   float64\n",
      " 10  sysBP            3658 non-null   float64\n",
      " 11  diaBP            3658 non-null   float64\n",
      " 12  BMI              3658 non-null   float64\n",
      " 13  heartRate        3658 non-null   float64\n",
      " 14  glucose          3658 non-null   float64\n",
      " 15  TenYearCHD       3658 non-null   int64  \n",
      "dtypes: float64(9), int64(7)\n",
      "memory usage: 485.8 KB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# Read the Framingham dataset\n",
    "framinghamEd = pd.read_csv(r\"D:\\Academics\\BTech\\DSC4.51015_StatisticalModeling-1\\Session 06\\framingham.csv\")\n",
    "# Remove rows with missing values from the DataFrame\n",
    "framinghamEd = framinghamEd.dropna()\n",
    "print(framinghamEd.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "427b5032",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 3658 entries, 0 to 4239\n",
      "Data columns (total 18 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   male             3658 non-null   int64  \n",
      " 1   age              3658 non-null   int64  \n",
      " 2   currentSmoker    3658 non-null   int64  \n",
      " 3   cigsPerDay       3658 non-null   float64\n",
      " 4   BPMeds           3658 non-null   float64\n",
      " 5   prevalentStroke  3658 non-null   int64  \n",
      " 6   prevalentHyp     3658 non-null   int64  \n",
      " 7   diabetes         3658 non-null   int64  \n",
      " 8   totChol          3658 non-null   float64\n",
      " 9   sysBP            3658 non-null   float64\n",
      " 10  diaBP            3658 non-null   float64\n",
      " 11  BMI              3658 non-null   float64\n",
      " 12  heartRate        3658 non-null   float64\n",
      " 13  glucose          3658 non-null   float64\n",
      " 14  TenYearCHD       3658 non-null   int64  \n",
      " 15  education_2.0    3658 non-null   bool   \n",
      " 16  education_3.0    3658 non-null   bool   \n",
      " 17  education_4.0    3658 non-null   bool   \n",
      "dtypes: bool(3), float64(8), int64(7)\n",
      "memory usage: 468.0 KB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# Convert 'education' to categorical\n",
    "framinghamEd['education'] = framinghamEd['education'].astype('category')\n",
    "# Create dummy variables for 'education'\n",
    "education_dummies = pd.get_dummies(framinghamEd['education'], prefix='education', drop_first=True)\n",
    "# Concatenate dummy variables with the original DataFrame\n",
    "framinghamEd = pd.concat([framinghamEd, education_dummies], axis=1)\n",
    "# Remove the original 'education' column if needed\n",
    "framinghamEd = framinghamEd.drop(columns=['education'])\n",
    "print(framinghamEd.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "435884c1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   male  age  currentSmoker  cigsPerDay  BPMeds  prevalentStroke  \\\n",
      "0     1   39              0         0.0     0.0                0   \n",
      "1     0   46              0         0.0     0.0                0   \n",
      "2     1   48              1        20.0     0.0                0   \n",
      "3     0   61              1        30.0     0.0                0   \n",
      "4     0   46              1        23.0     0.0                0   \n",
      "\n",
      "   prevalentHyp  diabetes  totChol  sysBP  diaBP    BMI  heartRate  glucose  \\\n",
      "0             0         0    195.0  106.0   70.0  26.97       80.0     77.0   \n",
      "1             0         0    250.0  121.0   81.0  28.73       95.0     76.0   \n",
      "2             0         0    245.0  127.5   80.0  25.34       75.0     70.0   \n",
      "3             1         0    225.0  150.0   95.0  28.58       65.0    103.0   \n",
      "4             0         0    285.0  130.0   84.0  23.10       85.0     85.0   \n",
      "\n",
      "   TenYearCHD  education_2.0  education_3.0  education_4.0  \n",
      "0           0          False          False           True  \n",
      "1           0           True          False          False  \n",
      "2           0          False          False          False  \n",
      "3           1          False           True          False  \n",
      "4           0          False           True          False  \n"
     ]
    }
   ],
   "source": [
    "print(framinghamEd.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "d206426c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert boolean columns to integer (0 or 1)\n",
    "framinghamEd['education_2.0'] = framinghamEd['education_2.0'].astype(int)\n",
    "framinghamEd['education_3.0'] = framinghamEd['education_3.0'].astype(int)\n",
    "framinghamEd['education_4.0'] = framinghamEd['education_4.0'].astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "55cb78f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "male               0\n",
      "age                0\n",
      "currentSmoker      0\n",
      "cigsPerDay         0\n",
      "BPMeds             0\n",
      "prevalentStroke    0\n",
      "prevalentHyp       0\n",
      "diabetes           0\n",
      "totChol            0\n",
      "sysBP              0\n",
      "diaBP              0\n",
      "BMI                0\n",
      "heartRate          0\n",
      "glucose            0\n",
      "TenYearCHD         0\n",
      "education_2.0      0\n",
      "education_3.0      0\n",
      "education_4.0      0\n",
      "dtype: int64\n",
      "male                 int64\n",
      "age                  int64\n",
      "currentSmoker        int64\n",
      "cigsPerDay         float64\n",
      "BPMeds             float64\n",
      "prevalentStroke      int64\n",
      "prevalentHyp         int64\n",
      "diabetes             int64\n",
      "totChol            float64\n",
      "sysBP              float64\n",
      "diaBP              float64\n",
      "BMI                float64\n",
      "heartRate          float64\n",
      "glucose            float64\n",
      "TenYearCHD           int64\n",
      "education_2.0        int32\n",
      "education_3.0        int32\n",
      "education_4.0        int32\n",
      "dtype: object\n"
     ]
    }
   ],
   "source": [
    "# Check for missing values\n",
    "print(framinghamEd.isnull().sum())\n",
    "\n",
    "# Verify data types\n",
    "print(framinghamEd.dtypes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "7d7f0db6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Randomly split the data into training and testing sets\n",
    "trainEd, testEd = train_test_split(framinghamEd, test_size=0.3, random_state=123)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "6cf18287",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Proportion of TenYearCHD in the training dataset: 0.147265625\n",
      "Proportion of TenYearCHD in the testing dataset: 0.16393442622950818\n"
     ]
    }
   ],
   "source": [
    "# Calculate the proportion of TenYearCHD in the training dataset\n",
    "trainEd_proportion = trainEd['TenYearCHD'].sum() / len(trainEd)\n",
    "\n",
    "# Calculate the proportion of TenYearCHD in the testing dataset\n",
    "testEd_proportion = testEd['TenYearCHD'].sum() / len(testEd)\n",
    "\n",
    "print(\"Proportion of TenYearCHD in the training dataset:\", trainEd_proportion)\n",
    "print(\"Proportion of TenYearCHD in the testing dataset:\", testEd_proportion)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "3150e260",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                Results: Generalized linear model\n",
      "=================================================================\n",
      "Model:               GLM              AIC:            1921.9700  \n",
      "Link Function:       Logit            BIC:            -18063.0423\n",
      "Dependent Variable:  TenYearCHD       Log-Likelihood: -942.99    \n",
      "Date:                2024-03-27 14:25 LL-Null:        -1069.9    \n",
      "No. Observations:    2560             Deviance:       1886.0     \n",
      "Df Model:            17               Pearson chi2:   2.46e+03   \n",
      "Df Residuals:        2542             Scale:          1.0000     \n",
      "Method:              IRLS                                        \n",
      "-----------------------------------------------------------------\n",
      "                 Coef.  Std.Err.    z     P>|z|   [0.025   0.975]\n",
      "-----------------------------------------------------------------\n",
      "const           -8.8713   0.8597 -10.3187 0.0000 -10.5563 -7.1863\n",
      "male             0.5142   0.1336   3.8500 0.0001   0.2525  0.7760\n",
      "age              0.0636   0.0081   7.8096 0.0000   0.0476  0.0795\n",
      "currentSmoker    0.1502   0.1905   0.7883 0.4305  -0.2232  0.5236\n",
      "cigsPerDay       0.0211   0.0077   2.7413 0.0061   0.0060  0.0363\n",
      "BPMeds           0.2804   0.2878   0.9745 0.3298  -0.2836  0.8445\n",
      "prevalentStroke  0.8437   0.5470   1.5426 0.1229  -0.2283  1.9158\n",
      "prevalentHyp    -0.0436   0.1696  -0.2572 0.7970  -0.3759  0.2887\n",
      "diabetes        -0.0099   0.3994  -0.0248 0.9802  -0.7926  0.7728\n",
      "totChol          0.0034   0.0014   2.4235 0.0154   0.0006  0.0061\n",
      "sysBP            0.0176   0.0047   3.7681 0.0002   0.0084  0.0267\n",
      "diaBP           -0.0040   0.0078  -0.5079 0.6115  -0.0193  0.0114\n",
      "BMI              0.0125   0.0155   0.8022 0.4224  -0.0180  0.0429\n",
      "heartRate       -0.0039   0.0051  -0.7632 0.4454  -0.0140  0.0061\n",
      "glucose          0.0060   0.0026   2.2934 0.0218   0.0009  0.0112\n",
      "education_2.0   -0.2388   0.1509  -1.5821 0.1136  -0.5347  0.0570\n",
      "education_3.0   -0.1630   0.1749  -0.9319 0.3514  -0.5058  0.1798\n",
      "education_4.0   -0.1759   0.2081  -0.8456 0.3978  -0.5837  0.2319\n",
      "=================================================================\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Logistic Regression Model\n",
    "framingham2 = sm.GLM(trainEd['TenYearCHD'], sm.add_constant(trainEd.drop(columns=['TenYearCHD'])), \n",
    "                     family=sm.families.Binomial()).fit()\n",
    "print(framingham2.summary2())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "9ef89c42",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "VIF:\n",
      "     Variable       VIF\n",
      "0         age  1.315494\n",
      "1  cigsPerDay  1.048542\n",
      "2     totChol  1.106154\n",
      "3       sysBP  3.109630\n",
      "4       diaBP  2.860657\n",
      "5         BMI  1.190182\n",
      "6   heartRate  1.062395\n",
      "7     glucose  1.038703\n"
     ]
    }
   ],
   "source": [
    "# Check for multicollinearity\n",
    "\n",
    "X = framinghamEd.drop(columns=['TenYearCHD']+['male']+['currentSmoker']+['BPMeds']+['prevalentStroke']+['prevalentHyp']\n",
    "                      +['diabetes']+['education_2.0']+['education_3.0']+['education_4.0'], axis=1)\n",
    "# Add a constant term for the intercept\n",
    "X_with_constant = sm.add_constant(X)\n",
    "# Calculate VIF (excluding the constant term)\n",
    "vif_data = pd.DataFrame()\n",
    "vif_data[\"Variable\"] = X.columns\n",
    "vif_data[\"VIF\"] = [variance_inflation_factor(X_with_constant.values, i) for i in range(1, X_with_constant.shape[1])]\n",
    "\n",
    "print(\"VIF:\")\n",
    "print(vif_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "16b52d65",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix_Train:\n",
      "[[2168   15]\n",
      " [ 346   31]]\n",
      "Accuracy_Train: 0.858984375\n",
      "Confusion Matrix_Test:\n",
      "[[913   5]\n",
      " [166  14]]\n",
      "Accuracy_Test: 0.8442622950819673\n"
     ]
    }
   ],
   "source": [
    "# Accuracy on the training set\n",
    "predictTrain2 = framingham2.predict(sm.add_constant(trainEd.drop(columns=['TenYearCHD'])))\n",
    "print(\"Confusion Matrix_Train:\")\n",
    "print(confusion_matrix(trainEd['TenYearCHD'], predictTrain2 > 0.5))\n",
    "print(\"Accuracy_Train:\", accuracy_score(trainEd['TenYearCHD'], predictTrain2 > 0.5))\n",
    "\n",
    "# Accuracy on Test Set\n",
    "predictTest2 = framingham2.predict(sm.add_constant(testEd.drop(columns=['TenYearCHD'])))\n",
    "print(\"Confusion Matrix_Test:\")\n",
    "print(confusion_matrix(testEd['TenYearCHD'], predictTest2 > 0.5))\n",
    "print(\"Accuracy_Test:\", accuracy_score(testEd['TenYearCHD'], predictTest2 > 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "99e3cb41",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7260288065843621\n",
      "0.7250847252481238\n"
     ]
    }
   ],
   "source": [
    "# Test set AUC using FULL MODEL (MODEL 1)\n",
    "print(roc_auc_score(test['TenYearCHD'], predictTest1))\n",
    "\n",
    "# Test set AUC using REDUCED MODEL (MODEL 2)\n",
    "print(roc_auc_score(test['TenYearCHD'], predictTest2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "ef6f1c2b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, roc_auc_score\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "7acf54b7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7260288065843621\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7250847252481238\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Test set AUC using MODEL 1\n",
    "auc1 = roc_auc_score(test['TenYearCHD'], predictTest1)\n",
    "print(auc1)\n",
    "\n",
    "fpr1, tpr1, thresholds1 = roc_curve(test['TenYearCHD'], predictTest1)\n",
    "\n",
    "plt.plot(fpr1, tpr1, color='blue', lw=2, label='Model1 ROC curve (AUC = %0.2f)' % auc1)\n",
    "plt.plot([0, 1], [0, 1], color='gray', linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('Receiver Operating Characteristic (Model1)')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "\n",
    "# Test set AUC using MODEL 2\n",
    "auc2 = roc_auc_score(testEd['TenYearCHD'], predictTest2)\n",
    "print(auc2)\n",
    "\n",
    "fpr2, tpr2, thresholds2 = roc_curve(testEd['TenYearCHD'], predictTest2)\n",
    "\n",
    "plt.plot(fpr2, tpr2, color='green', lw=2, label='Model2 ROC curve (AUC = %0.2f)' % auc2)\n",
    "plt.plot([0, 1], [0, 1], color='gray', linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('Receiver Operating Characteristic (Model2)')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "\n",
    "# Overlay ROC curves\n",
    "plt.plot(fpr1, tpr1, color='blue', lw=2, label='Model1 ROC curve (AUC = %0.2f)' % auc1)\n",
    "plt.plot(fpr2, tpr2, color='green', lw=2, label='Model2 ROC curve (AUC = %0.2f)' % auc2)\n",
    "plt.plot([0, 1], [0, 1], color='gray', linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "f9c82ec7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      Predicted Probability  Actual Response\n",
      "3844               0.922906                1\n",
      "2007               0.672367                1\n",
      "2718               0.637595                1\n",
      "1588               0.613555                1\n",
      "2930               0.609554                1\n",
      "590                0.576604                0\n",
      "604                0.575652                1\n",
      "249                0.557458                0\n",
      "1230               0.557357                1\n",
      "1054               0.549943                0\n",
      "1079               0.546651                1\n",
      "2162               0.539469                0\n",
      "3136               0.531380                1\n",
      "2187               0.530724                1\n",
      "4075               0.530574                1\n",
      "1878               0.515399                1\n",
      "2600               0.512034                1\n",
      "926                0.506415                0\n",
      "2298               0.505908                1\n",
      "2446               0.488357                0\n",
      "4040               0.482958                0\n",
      "522                0.472157                1\n",
      "1485               0.471319                1\n",
      "153                0.470231                1\n",
      "1399               0.466439                0\n"
     ]
    }
   ],
   "source": [
    "# Combine predicted probabilities and actual response\n",
    "probs_and_actual = pd.DataFrame({'Predicted Probability': predictTest1, 'Actual Response': test['TenYearCHD']})\n",
    "\n",
    "# Sort by predicted probability in descending order\n",
    "probs_and_actual_sorted = probs_and_actual.sort_values(by='Predicted Probability', ascending=False)\n",
    "\n",
    "# Print the first 25 records\n",
    "print(probs_and_actual_sorted.head(25))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b129810b",
   "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": "3.12.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}