FinalProject_group050.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# COGS 118A Project - Heart Disease Detection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Names\n",
    "- Zhaoge Ouyang\n",
    "- Bateliel Bassa\n",
    "- Jason Hsu\n",
    "- Kevin Thai\n",
    "- Katelyn Dang"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Abstract \n",
    "Heart disease is one of the leading medical issues in the U.S. that has affected “as of 2018, 30.3 million U.S. adults”<a name=\"https://www.healthline.com/health/heart-disease/statistics\"></a>[<sup>[1]</sup>](https://www.healthline.com/health/heart-disease/statistics) as reported by the CDC (Centers for Disease Control and Prevention). Our team would like to find out patterns that may lead to high risk of heart disease and possibly predict if a patient is at a high risk. We will be using 2020 annual CDC survey data of 320k adults related to their health status, provided on Kaggle. The data includes various variables regarding a patient's lifestyle and physical health, such as a person's BMI or whether they have smoked a certain number of cigarettes in their life, measured as a yes/no question. We will be training models to detect “patterns”, such that we can predict a patient’s heart condition and whether or not a patient is more likely to have heart diseases. We will split the dataset to 80% training, 10% validation, and 10% test set. With around 320,000 valid observations, the training set will have about 256,000 observations, while the validation and test set will have 64,000 observations each. The performance will be measured with accuracy."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Background\n",
    "\n",
    "Heart disease is one of the leading causes of death around the world. According to CDC, 659,000 people die from heart diseases in the United States. There are many factors that contribute to heart disease, including lifestyle and genetics. Health conditions such as high blood pressure, unhealthy blood cholesterol level, diabetes, and obesity increase the risk of heart dieseases. In addition, behaviors such as a diet high in saturated fat, trans fat, lack of physical exercise, alcohol abuse, tobacco use also increases a patient’s risk of heart diseases. Other than life styles, genetics and race also affect the risk of heart diseases. Genetic factors affect the risk of health conditions that are related to heart diseases, and some races are more likely to to be at risk of heart diseases than others. Lastly, heart diseases tend to happen more often as age goes up<a name=\"admonish\"></a>[<sup>[2]</sup>](#admonishnote)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Problem Statement\n",
    "\n",
    "Currently, there aren’t many tools that are able to predict the likelihood of whether a patient is at risk of heart diseases based on their lifestyle and genetics. Therefore, we want to train a model that can properly make predictions on whether a patient may have heart disease.\n",
    " Whether the patient is at potential risk of heart disease will be measured as a simple 0 or 1 output, 0 for if the patient is less likely to be at risk and 1 for if the patient is at possible risk. By reviewing the prediction outcomes, we will also be able to determine which parameters contribute the most heavily to potential risk of heart diseases. Given appropriate data that matches the variables/parameters our model is trained on, the model should be able to produce a prediction of a patient’s risk for heart disease. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data\n",
    "\n",
    "We will be using 2020 annual CDC survey data of 320k adults related to their health status, obtained on Kaggle<a name=\"https://www.kaggle.com/datasets/kamilpytlak/personal-key-indicators-of-heart-disease\"></a>[<sup>[3]</sup>](https://www.kaggle.com/datasets/kamilpytlak/personal-key-indicators-of-heart-disease). The dataset includes 18 variables and ~320,000 observations. Variables includes health related status such as had heart disease, smoked 100 cigarettes in their lifetime, having at least 7 (female) /14 (male) drinks a week depending, had a stroke, considered their physical or mental health to be “unwell” over the past 30 days, difficulty walking up stairs, gender, age, and BMI.\n",
    "\n",
    "The dataset includes various kinds of data types.Gender questions were answered as Male/Female, race was measured by the participants ethnicity, General health was measured on a likert scale, age category was answered as a range from 18-24 to 80-over, BMI and sleeping time was measured as a specific number, physical and mental health from a range from 0-30, and all other variables were answered as Yes/No.From the dataset, the data on heart diseases, smoking, age, physical activity, and race are the most important to help answer our research question. Smoking, age, physical activity, and race are suspected links to heart diseases and by studying and analyzing the data between heart diseases and these links, we may be able to find which variables can increase the risk of heart disease. \n",
    "\n",
    "To better handle the data, we will need to transform binary data types and categorical data into one-hot encoding (i.e. The Gender column would be split into Male and Female columns). In addition, there are only about 9% of data that are labeled as having heart disease or myocardial infarction. To combat the unbalanced proportion of data is to undersample our data (reduce the number of data that contains negative cases). Note that if we proceed with undersampling data, we will lose a lot of information. However, since the dataset is quite huge, we should still have a sufficient amount of data to work with. Another possible problem we may encounter is that many features seem to also be unbalanced (i.e. proportion of people who have kidney disease is much lower than people who don’t). Therefore, we may also need to balance weights by giving the smaller class more weights.\n",
    "\n",
    "\n",
    "There are certain features of this dataset that may be particularly useful in attempting to predict risk for heart disease. Some of these features are smoking, drinking alcohol, physical health, mental health, age, physical activity, as well as sex and general health. Our reasoning for choosing these features as the most important is that they come from the CDC website regarding risks for heart disease<a name=\"https://www.cdc.gov/heartdisease/risk_factors.htm\"></a>[<sup>[4]</sup>](https://www.cdc.gov/heartdisease/risk_factors.htm). These features also line up with what we would think are relevant factors to heart disease, especially features such as smoking and physical health and activity.\n",
    "\n",
    "### Data Handling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HeartDisease</th>\n",
       "      <th>BMI</th>\n",
       "      <th>Smoking</th>\n",
       "      <th>AlcoholDrinking</th>\n",
       "      <th>Stroke</th>\n",
       "      <th>PhysicalHealth</th>\n",
       "      <th>MentalHealth</th>\n",
       "      <th>DiffWalking</th>\n",
       "      <th>Sex</th>\n",
       "      <th>AgeCategory</th>\n",
       "      <th>Race</th>\n",
       "      <th>Diabetic</th>\n",
       "      <th>PhysicalActivity</th>\n",
       "      <th>GenHealth</th>\n",
       "      <th>SleepTime</th>\n",
       "      <th>Asthma</th>\n",
       "      <th>KidneyDisease</th>\n",
       "      <th>SkinCancer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>No</td>\n",
       "      <td>16.60</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>3.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Female</td>\n",
       "      <td>55-59</td>\n",
       "      <td>White</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Very good</td>\n",
       "      <td>5.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>No</td>\n",
       "      <td>20.34</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Female</td>\n",
       "      <td>80 or older</td>\n",
       "      <td>White</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Very good</td>\n",
       "      <td>7.0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>No</td>\n",
       "      <td>26.58</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>20.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Male</td>\n",
       "      <td>65-69</td>\n",
       "      <td>White</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Fair</td>\n",
       "      <td>8.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>No</td>\n",
       "      <td>24.21</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Female</td>\n",
       "      <td>75-79</td>\n",
       "      <td>White</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Good</td>\n",
       "      <td>6.0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>No</td>\n",
       "      <td>23.71</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>28.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Female</td>\n",
       "      <td>40-44</td>\n",
       "      <td>White</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Very good</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  HeartDisease    BMI Smoking AlcoholDrinking Stroke  PhysicalHealth  \\\n",
       "0           No  16.60     Yes              No     No             3.0   \n",
       "1           No  20.34      No              No    Yes             0.0   \n",
       "2           No  26.58     Yes              No     No            20.0   \n",
       "3           No  24.21      No              No     No             0.0   \n",
       "4           No  23.71      No              No     No            28.0   \n",
       "\n",
       "   MentalHealth DiffWalking     Sex  AgeCategory   Race Diabetic  \\\n",
       "0          30.0          No  Female        55-59  White      Yes   \n",
       "1           0.0          No  Female  80 or older  White       No   \n",
       "2          30.0          No    Male        65-69  White      Yes   \n",
       "3           0.0          No  Female        75-79  White       No   \n",
       "4           0.0         Yes  Female        40-44  White       No   \n",
       "\n",
       "  PhysicalActivity  GenHealth  SleepTime Asthma KidneyDisease SkinCancer  \n",
       "0              Yes  Very good        5.0    Yes            No        Yes  \n",
       "1              Yes  Very good        7.0     No            No         No  \n",
       "2              Yes       Fair        8.0    Yes            No         No  \n",
       "3               No       Good        6.0     No            No        Yes  \n",
       "4              Yes  Very good        8.0     No            No         No  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('heart_2020_cleaned.csv')\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Columns:Index(['HeartDisease', 'BMI', 'Smoking', 'AlcoholDrinking', 'Stroke',\n",
      "       'PhysicalHealth', 'MentalHealth', 'DiffWalking', 'Sex', 'AgeCategory',\n",
      "       'Race', 'Diabetic', 'PhysicalActivity', 'GenHealth', 'SleepTime',\n",
      "       'Asthma', 'KidneyDisease', 'SkinCancer'],\n",
      "      dtype='object')\n",
      "Shapes:(319795, 18)\n"
     ]
    }
   ],
   "source": [
    "print(f'Columns:{df.columns}\\nShapes:{df.shape}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Shapes shown above shows that there are 319795 # of participants, 18 columns (17 variables with 1 label we are trying to predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Look at the unique values form each catagorical data and decide how to conver the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Catagorical Data Unique Outputs\n",
      "Heart Disease: ['No' 'Yes']\n",
      "Smoking: ['Yes' 'No']\n",
      "Alcohol Drinking: ['No' 'Yes']\n",
      "Stroke: ['No' 'Yes']\n",
      "Difficult in walking: ['No' 'Yes']\n",
      "Sex: ['Female' 'Male']\n",
      "Age: ['55-59' '80 or older' '65-69' '75-79' '40-44' '70-74' '60-64' '50-54'\n",
      " '45-49' '18-24' '35-39' '30-34' '25-29']\n",
      "Race: ['White' 'Black' 'Asian' 'American Indian/Alaskan Native' 'Other'\n",
      " 'Hispanic']\n",
      "Diabetic: ['Yes' 'No' 'No, borderline diabetes' 'Yes (during pregnancy)']\n",
      "Physical Activity: ['Yes' 'No']\n",
      "General Health: ['Very good' 'Fair' 'Good' 'Poor' 'Excellent']\n",
      "Asthma: ['Yes' 'No']\n",
      "Kidney Disease: ['No' 'Yes']\n",
      "Skin Cancer: ['Yes' 'No']\n"
     ]
    }
   ],
   "source": [
    "print(\"Catagorical Data Unique Outputs\")\n",
    "print(f\"Heart Disease: {df['HeartDisease'].unique()}\")\n",
    "print(f\"Smoking: {df['Smoking'].unique()}\")\n",
    "print(f\"Alcohol Drinking: {df['AlcoholDrinking'].unique()}\")\n",
    "print(f\"Stroke: {df['Stroke'].unique()}\")\n",
    "print(f\"Difficult in walking: {df['DiffWalking'].unique()}\")\n",
    "print(f\"Sex: {df['Sex'].unique()}\")\n",
    "print(f\"Age: {df['AgeCategory'].unique()}\")\n",
    "print(f\"Race: {df['Race'].unique()}\")\n",
    "print(f\"Diabetic: {df['Diabetic'].unique()}\")\n",
    "print(f\"Physical Activity: {df['PhysicalActivity'].unique()}\")\n",
    "print(f\"General Health: {df['GenHealth'].unique()}\")\n",
    "print(f\"Asthma: {df['Asthma'].unique()}\")\n",
    "print(f\"Kidney Disease: {df['KidneyDisease'].unique()}\")\n",
    "print(f\"Skin Cancer: {df['SkinCancer'].unique()}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Encode Guide\n",
    "- Yes/No categories (multiple categories) => binary (1/0)\n",
    "- Race => one hot encoding\n",
    "- Diabetic => one hot encoding\n",
    "- Age => int rank (order encoded)\n",
    "- Sex => binary (renamed to Male as in Male = 1, Female = 0)\n",
    "- General Health category => int rank (order encoded)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Functions\n",
    "yes_no_to_binary = lambda elem : 1 if elem == 'Yes' else 0\n",
    "sex_to_binary = lambda elem : 1 if elem == 'Male' else 0\n",
    "def genhealth_to_rank(elem):\n",
    "    if elem == 'Excellent':\n",
    "        return 4\n",
    "    if elem == 'Very good':\n",
    "        return 3\n",
    "    if elem == 'Good':\n",
    "        return 2\n",
    "    if elem == 'Fair':\n",
    "        return 1\n",
    "    if elem == 'Poor':\n",
    "        return 0\n",
    "    return \"ERROR\"\n",
    "\n",
    "def encode_age(age):\n",
    "    if age == '18-24':\n",
    "        return 18\n",
    "    elif age == '25-29':\n",
    "        return 25\n",
    "    elif age == '30-34':\n",
    "        return 30\n",
    "    elif age == '35-39':\n",
    "        return 35\n",
    "    elif age == '40-44':\n",
    "        return 40\n",
    "    elif age == '45-49':\n",
    "        return 45\n",
    "    elif age == '50-54':\n",
    "        return 50\n",
    "    elif age == '55-59':\n",
    "        return 55\n",
    "    elif age == '60-64':\n",
    "        return 60\n",
    "    elif age == '65-69':\n",
    "        return 65\n",
    "    elif age == '70-74':\n",
    "        return 70\n",
    "    elif age == '75-79':\n",
    "        return 75\n",
    "    return 80"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Yes/No Convertion\n",
    "df['HeartDisease'] = df['HeartDisease'].apply(yes_no_to_binary)\n",
    "df['Smoking'] = df['Smoking'].apply(yes_no_to_binary)\n",
    "df['AlcoholDrinking'] = df['AlcoholDrinking'].apply(yes_no_to_binary)\n",
    "df['Stroke'] = df['Stroke'].apply(yes_no_to_binary)\n",
    "df['DiffWalking'] = df['DiffWalking'].apply(yes_no_to_binary)\n",
    "df['PhysicalActivity'] = df['PhysicalActivity'].apply(yes_no_to_binary)\n",
    "df['Asthma'] = df['Asthma'].apply(yes_no_to_binary)\n",
    "df['KidneyDisease'] = df['KidneyDisease'].apply(yes_no_to_binary)\n",
    "df['SkinCancer'] = df['SkinCancer'].apply(yes_no_to_binary)\n",
    "\n",
    "# Dummy Variables Convertion\n",
    "dummy_race = pd.get_dummies(df['Race'])\n",
    "dummy_race.rename(columns={'American Indian/Alaskan Native':'American_Indian/Alaskan_Native'}, inplace=True)\n",
    "dummy_race.rename(columns={'Other':'Other_Race'}, inplace=True)\n",
    "df = pd.concat([df, dummy_race], axis=1).drop('Race', axis=1)\n",
    "\n",
    "dummy_diabetic = pd.get_dummies(df['Diabetic'])\n",
    "dummy_diabetic.rename(columns={'No, borderline diabetes':'Borderline_Diabetic'}, inplace=True)\n",
    "dummy_diabetic.rename(columns={'Yes (during pregnancy)':'Diabetic_during_pregnancy'}, inplace=True)\n",
    "dummy_diabetic.rename(columns={'Yes':'Diabetic'}, inplace=True)\n",
    "dummy_diabetic.rename(columns={'No':'Not_Diabetic'}, inplace=True)\n",
    "df = pd.concat([df, dummy_diabetic], axis=1).drop('Diabetic', axis=1)\n",
    "\n",
    "# Age Convertion\n",
    "df.rename(columns={'AgeCategory':'Age'}, inplace=True)\n",
    "df['Age'] = df['Age'].apply(encode_age)\n",
    "\n",
    "# Sex Convertion\n",
    "df.rename(columns={'Sex':'Male'}, inplace=True)\n",
    "df['Male'] = df['Male'].apply(sex_to_binary)\n",
    "\n",
    "# General Health Convertion\n",
    "df['GenHealth'] = df['GenHealth'].apply(genhealth_to_rank)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 319795 entries, 0 to 319794\n",
      "Data columns (total 25 columns):\n",
      " #   Column                          Non-Null Count   Dtype  \n",
      "---  ------                          --------------   -----  \n",
      " 0   HeartDisease                    319795 non-null  int64  \n",
      " 1   BMI                             319795 non-null  float64\n",
      " 2   Smoking                         319795 non-null  int64  \n",
      " 3   AlcoholDrinking                 319795 non-null  int64  \n",
      " 4   Stroke                          319795 non-null  int64  \n",
      " 5   PhysicalHealth                  319795 non-null  float64\n",
      " 6   MentalHealth                    319795 non-null  float64\n",
      " 7   DiffWalking                     319795 non-null  int64  \n",
      " 8   Male                            319795 non-null  int64  \n",
      " 9   Age                             319795 non-null  int64  \n",
      " 10  PhysicalActivity                319795 non-null  int64  \n",
      " 11  GenHealth                       319795 non-null  int64  \n",
      " 12  SleepTime                       319795 non-null  float64\n",
      " 13  Asthma                          319795 non-null  int64  \n",
      " 14  KidneyDisease                   319795 non-null  int64  \n",
      " 15  SkinCancer                      319795 non-null  int64  \n",
      " 16  American_Indian/Alaskan_Native  319795 non-null  uint8  \n",
      " 17  Asian                           319795 non-null  uint8  \n",
      " 18  Black                           319795 non-null  uint8  \n",
      " 19  Hispanic                        319795 non-null  uint8  \n",
      " 20  Other_Race                      319795 non-null  uint8  \n",
      " 21  White                           319795 non-null  uint8  \n",
      " 22  Not_Diabetic                    319795 non-null  uint8  \n",
      " 23  Borderline_Diabetic             319795 non-null  uint8  \n",
      " 24  Diabetic_during_pregnancy       319795 non-null  uint8  \n",
      "dtypes: float64(4), int64(12), uint8(9)\n",
      "memory usage: 41.8 MB\n"
     ]
    }
   ],
   "source": [
    "# Show datatype of cleaned data\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1080 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# View the distribution of the data\n",
    "df.hist(figsize=(20,15))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As shown above, the amount of people who does not have heart disease significantly outweights people who have."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>HeartDisease</th>\n",
       "      <th>BMI</th>\n",
       "      <th>Smoking</th>\n",
       "      <th>AlcoholDrinking</th>\n",
       "      <th>Stroke</th>\n",
       "      <th>PhysicalHealth</th>\n",
       "      <th>MentalHealth</th>\n",
       "      <th>DiffWalking</th>\n",
       "      <th>Male</th>\n",
       "      <th>Age</th>\n",
       "      <th>...</th>\n",
       "      <th>SkinCancer</th>\n",
       "      <th>American_Indian/Alaskan_Native</th>\n",
       "      <th>Asian</th>\n",
       "      <th>Black</th>\n",
       "      <th>Hispanic</th>\n",
       "      <th>Other_Race</th>\n",
       "      <th>White</th>\n",
       "      <th>Not_Diabetic</th>\n",
       "      <th>Borderline_Diabetic</th>\n",
       "      <th>Diabetic_during_pregnancy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>HeartDisease</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.051803</td>\n",
       "      <td>0.107764</td>\n",
       "      <td>-0.032080</td>\n",
       "      <td>0.196835</td>\n",
       "      <td>0.170721</td>\n",
       "      <td>0.028591</td>\n",
       "      <td>0.201258</td>\n",
       "      <td>0.070040</td>\n",
       "      <td>0.232325</td>\n",
       "      <td>...</td>\n",
       "      <td>0.093317</td>\n",
       "      <td>0.008547</td>\n",
       "      <td>-0.030262</td>\n",
       "      <td>-0.010156</td>\n",
       "      <td>-0.036163</td>\n",
       "      <td>-0.003039</td>\n",
       "      <td>0.040121</td>\n",
       "      <td>-0.170977</td>\n",
       "      <td>0.016182</td>\n",
       "      <td>-0.013930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BMI</th>\n",
       "      <td>0.051803</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.023118</td>\n",
       "      <td>-0.038816</td>\n",
       "      <td>0.019733</td>\n",
       "      <td>0.109788</td>\n",
       "      <td>0.064131</td>\n",
       "      <td>0.181678</td>\n",
       "      <td>0.026940</td>\n",
       "      <td>-0.001740</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.033644</td>\n",
       "      <td>0.026347</td>\n",
       "      <td>-0.078643</td>\n",
       "      <td>0.080672</td>\n",
       "      <td>0.023750</td>\n",
       "      <td>0.012535</td>\n",
       "      <td>-0.049060</td>\n",
       "      <td>-0.208232</td>\n",
       "      <td>0.049231</td>\n",
       "      <td>0.008198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Smoking</th>\n",
       "      <td>0.107764</td>\n",
       "      <td>0.023118</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.111768</td>\n",
       "      <td>0.061226</td>\n",
       "      <td>0.115352</td>\n",
       "      <td>0.085157</td>\n",
       "      <td>0.120074</td>\n",
       "      <td>0.085052</td>\n",
       "      <td>0.130384</td>\n",
       "      <td>...</td>\n",
       "      <td>0.033977</td>\n",
       "      <td>0.035667</td>\n",
       "      <td>-0.060308</td>\n",
       "      <td>-0.038249</td>\n",
       "      <td>-0.066873</td>\n",
       "      <td>0.016205</td>\n",
       "      <td>0.072367</td>\n",
       "      <td>-0.055127</td>\n",
       "      <td>0.007143</td>\n",
       "      <td>-0.005385</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AlcoholDrinking</th>\n",
       "      <td>-0.032080</td>\n",
       "      <td>-0.038816</td>\n",
       "      <td>0.111768</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.019858</td>\n",
       "      <td>-0.017254</td>\n",
       "      <td>0.051282</td>\n",
       "      <td>-0.035328</td>\n",
       "      <td>0.004200</td>\n",
       "      <td>-0.058837</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.005702</td>\n",
       "      <td>-0.004243</td>\n",
       "      <td>-0.022275</td>\n",
       "      <td>-0.025980</td>\n",
       "      <td>-0.015554</td>\n",
       "      <td>0.005455</td>\n",
       "      <td>0.033342</td>\n",
       "      <td>0.057644</td>\n",
       "      <td>-0.008165</td>\n",
       "      <td>-0.004077</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Stroke</th>\n",
       "      <td>0.196835</td>\n",
       "      <td>0.019733</td>\n",
       "      <td>0.061226</td>\n",
       "      <td>-0.019858</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.137014</td>\n",
       "      <td>0.046467</td>\n",
       "      <td>0.174143</td>\n",
       "      <td>-0.003091</td>\n",
       "      <td>0.137280</td>\n",
       "      <td>...</td>\n",
       "      <td>0.048116</td>\n",
       "      <td>0.014357</td>\n",
       "      <td>-0.016060</td>\n",
       "      <td>0.024819</td>\n",
       "      <td>-0.025823</td>\n",
       "      <td>0.005743</td>\n",
       "      <td>0.001154</td>\n",
       "      <td>-0.102787</td>\n",
       "      <td>0.011172</td>\n",
       "      <td>-0.004526</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PhysicalHealth</th>\n",
       "      <td>0.170721</td>\n",
       "      <td>0.109788</td>\n",
       "      <td>0.115352</td>\n",
       "      <td>-0.017254</td>\n",
       "      <td>0.137014</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.287987</td>\n",
       "      <td>0.428373</td>\n",
       "      <td>-0.040904</td>\n",
       "      <td>0.110789</td>\n",
       "      <td>...</td>\n",
       "      <td>0.041700</td>\n",
       "      <td>0.022955</td>\n",
       "      <td>-0.035229</td>\n",
       "      <td>0.011371</td>\n",
       "      <td>-0.005571</td>\n",
       "      <td>0.016102</td>\n",
       "      <td>-0.003968</td>\n",
       "      <td>-0.153469</td>\n",
       "      <td>0.021437</td>\n",
       "      <td>-0.000773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MentalHealth</th>\n",
       "      <td>0.028591</td>\n",
       "      <td>0.064131</td>\n",
       "      <td>0.085157</td>\n",
       "      <td>0.051282</td>\n",
       "      <td>0.046467</td>\n",
       "      <td>0.287987</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.152235</td>\n",
       "      <td>-0.100058</td>\n",
       "      <td>-0.155455</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.033412</td>\n",
       "      <td>0.018394</td>\n",
       "      <td>-0.023113</td>\n",
       "      <td>0.010944</td>\n",
       "      <td>0.011167</td>\n",
       "      <td>0.031385</td>\n",
       "      <td>-0.024489</td>\n",
       "      <td>-0.032353</td>\n",
       "      <td>0.011056</td>\n",
       "      <td>0.018742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DiffWalking</th>\n",
       "      <td>0.201258</td>\n",
       "      <td>0.181678</td>\n",
       "      <td>0.120074</td>\n",
       "      <td>-0.035328</td>\n",
       "      <td>0.174143</td>\n",
       "      <td>0.428373</td>\n",
       "      <td>0.152235</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.068860</td>\n",
       "      <td>0.242552</td>\n",
       "      <td>...</td>\n",
       "      <td>0.064840</td>\n",
       "      <td>0.025205</td>\n",
       "      <td>-0.038830</td>\n",
       "      <td>0.039615</td>\n",
       "      <td>-0.008214</td>\n",
       "      <td>0.008483</td>\n",
       "      <td>-0.015516</td>\n",
       "      <td>-0.209640</td>\n",
       "      <td>0.031843</td>\n",
       "      <td>-0.007040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Male</th>\n",
       "      <td>0.070040</td>\n",
       "      <td>0.026940</td>\n",
       "      <td>0.085052</td>\n",
       "      <td>0.004200</td>\n",
       "      <td>-0.003091</td>\n",
       "      <td>-0.040904</td>\n",
       "      <td>-0.100058</td>\n",
       "      <td>-0.068860</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.067682</td>\n",
       "      <td>...</td>\n",
       "      <td>0.013434</td>\n",
       "      <td>-0.003434</td>\n",
       "      <td>0.013716</td>\n",
       "      <td>-0.037280</td>\n",
       "      <td>-0.003607</td>\n",
       "      <td>0.008762</td>\n",
       "      <td>0.017314</td>\n",
       "      <td>0.003636</td>\n",
       "      <td>-0.004295</td>\n",
       "      <td>-0.085477</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Age</th>\n",
       "      <td>0.232325</td>\n",
       "      <td>-0.001740</td>\n",
       "      <td>0.130384</td>\n",
       "      <td>-0.058837</td>\n",
       "      <td>0.137280</td>\n",
       "      <td>0.110789</td>\n",
       "      <td>-0.155455</td>\n",
       "      <td>0.242552</td>\n",
       "      <td>-0.067682</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.262208</td>\n",
       "      <td>-0.024922</td>\n",
       "      <td>-0.084104</td>\n",
       "      <td>-0.034563</td>\n",
       "      <td>-0.162269</td>\n",
       "      <td>-0.061319</td>\n",
       "      <td>0.193558</td>\n",
       "      <td>-0.204576</td>\n",
       "      <td>0.048657</td>\n",
       "      <td>-0.033325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PhysicalActivity</th>\n",
       "      <td>-0.100030</td>\n",
       "      <td>-0.150616</td>\n",
       "      <td>-0.097174</td>\n",
       "      <td>0.017487</td>\n",
       "      <td>-0.079455</td>\n",
       "      <td>-0.232283</td>\n",
       "      <td>-0.095808</td>\n",
       "      <td>-0.278524</td>\n",
       "      <td>0.048247</td>\n",
       "      <td>-0.121465</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.001328</td>\n",
       "      <td>-0.017736</td>\n",
       "      <td>0.012106</td>\n",
       "      <td>-0.043701</td>\n",
       "      <td>-0.056564</td>\n",
       "      <td>0.001808</td>\n",
       "      <td>0.064172</td>\n",
       "      <td>0.135451</td>\n",
       "      <td>-0.017253</td>\n",
       "      <td>0.001501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GenHealth</th>\n",
       "      <td>-0.243182</td>\n",
       "      <td>-0.230720</td>\n",
       "      <td>-0.174717</td>\n",
       "      <td>0.029780</td>\n",
       "      <td>-0.168090</td>\n",
       "      <td>-0.482697</td>\n",
       "      <td>-0.241625</td>\n",
       "      <td>-0.413797</td>\n",
       "      <td>0.025947</td>\n",
       "      <td>-0.188236</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.050641</td>\n",
       "      <td>-0.037600</td>\n",
       "      <td>0.032701</td>\n",
       "      <td>-0.062215</td>\n",
       "      <td>-0.028721</td>\n",
       "      <td>-0.015241</td>\n",
       "      <td>0.062656</td>\n",
       "      <td>0.273381</td>\n",
       "      <td>-0.047783</td>\n",
       "      <td>-0.001201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SleepTime</th>\n",
       "      <td>0.008327</td>\n",
       "      <td>-0.051822</td>\n",
       "      <td>-0.030336</td>\n",
       "      <td>-0.005065</td>\n",
       "      <td>0.011900</td>\n",
       "      <td>-0.061387</td>\n",
       "      <td>-0.119717</td>\n",
       "      <td>-0.022216</td>\n",
       "      <td>-0.015704</td>\n",
       "      <td>0.103071</td>\n",
       "      <td>...</td>\n",
       "      <td>0.041266</td>\n",
       "      <td>-0.003615</td>\n",
       "      <td>-0.019985</td>\n",
       "      <td>-0.020192</td>\n",
       "      <td>-0.013242</td>\n",
       "      <td>-0.031007</td>\n",
       "      <td>0.042905</td>\n",
       "      <td>-0.000494</td>\n",
       "      <td>-0.006957</td>\n",
       "      <td>-0.011668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Asthma</th>\n",
       "      <td>0.041444</td>\n",
       "      <td>0.092345</td>\n",
       "      <td>0.024149</td>\n",
       "      <td>-0.002202</td>\n",
       "      <td>0.038866</td>\n",
       "      <td>0.117907</td>\n",
       "      <td>0.114008</td>\n",
       "      <td>0.103222</td>\n",
       "      <td>-0.069191</td>\n",
       "      <td>-0.058235</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.000396</td>\n",
       "      <td>0.013757</td>\n",
       "      <td>-0.017007</td>\n",
       "      <td>0.021297</td>\n",
       "      <td>0.003885</td>\n",
       "      <td>0.028037</td>\n",
       "      <td>-0.025421</td>\n",
       "      <td>-0.050782</td>\n",
       "      <td>0.016750</td>\n",
       "      <td>0.013282</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KidneyDisease</th>\n",
       "      <td>0.145197</td>\n",
       "      <td>0.050768</td>\n",
       "      <td>0.034920</td>\n",
       "      <td>-0.028280</td>\n",
       "      <td>0.091167</td>\n",
       "      <td>0.142197</td>\n",
       "      <td>0.037281</td>\n",
       "      <td>0.153064</td>\n",
       "      <td>-0.009084</td>\n",
       "      <td>0.122697</td>\n",
       "      <td>...</td>\n",
       "      <td>0.061816</td>\n",
       "      <td>0.007401</td>\n",
       "      <td>-0.016957</td>\n",
       "      <td>0.010300</td>\n",
       "      <td>-0.010960</td>\n",
       "      <td>-0.002057</td>\n",
       "      <td>0.005933</td>\n",
       "      <td>-0.142371</td>\n",
       "      <td>0.004982</td>\n",
       "      <td>-0.004706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SkinCancer</th>\n",
       "      <td>0.093317</td>\n",
       "      <td>-0.033644</td>\n",
       "      <td>0.033977</td>\n",
       "      <td>-0.005702</td>\n",
       "      <td>0.048116</td>\n",
       "      <td>0.041700</td>\n",
       "      <td>-0.033412</td>\n",
       "      <td>0.064840</td>\n",
       "      <td>0.013434</td>\n",
       "      <td>0.262208</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.026784</td>\n",
       "      <td>-0.047749</td>\n",
       "      <td>-0.083390</td>\n",
       "      <td>-0.082360</td>\n",
       "      <td>-0.031905</td>\n",
       "      <td>0.144862</td>\n",
       "      <td>-0.034473</td>\n",
       "      <td>0.006025</td>\n",
       "      <td>-0.010092</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>American_Indian/Alaskan_Native</th>\n",
       "      <td>0.008547</td>\n",
       "      <td>0.026347</td>\n",
       "      <td>0.035667</td>\n",
       "      <td>-0.004243</td>\n",
       "      <td>0.014357</td>\n",
       "      <td>0.022955</td>\n",
       "      <td>0.018394</td>\n",
       "      <td>0.025205</td>\n",
       "      <td>-0.003434</td>\n",
       "      <td>-0.024922</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.026784</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.020687</td>\n",
       "      <td>-0.035746</td>\n",
       "      <td>-0.039400</td>\n",
       "      <td>-0.024188</td>\n",
       "      <td>-0.233164</td>\n",
       "      <td>-0.028442</td>\n",
       "      <td>0.008356</td>\n",
       "      <td>0.003433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Asian</th>\n",
       "      <td>-0.030262</td>\n",
       "      <td>-0.078643</td>\n",
       "      <td>-0.060308</td>\n",
       "      <td>-0.022275</td>\n",
       "      <td>-0.016060</td>\n",
       "      <td>-0.035229</td>\n",
       "      <td>-0.023113</td>\n",
       "      <td>-0.038830</td>\n",
       "      <td>0.013716</td>\n",
       "      <td>-0.084104</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.047749</td>\n",
       "      <td>-0.020687</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.044721</td>\n",
       "      <td>-0.049293</td>\n",
       "      <td>-0.030261</td>\n",
       "      <td>-0.291707</td>\n",
       "      <td>0.002578</td>\n",
       "      <td>0.023243</td>\n",
       "      <td>0.007260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Black</th>\n",
       "      <td>-0.010156</td>\n",
       "      <td>0.080672</td>\n",
       "      <td>-0.038249</td>\n",
       "      <td>-0.025980</td>\n",
       "      <td>0.024819</td>\n",
       "      <td>0.011371</td>\n",
       "      <td>0.010944</td>\n",
       "      <td>0.039615</td>\n",
       "      <td>-0.037280</td>\n",
       "      <td>-0.034563</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.083390</td>\n",
       "      <td>-0.035746</td>\n",
       "      <td>-0.044721</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.085173</td>\n",
       "      <td>-0.052288</td>\n",
       "      <td>-0.504040</td>\n",
       "      <td>-0.055268</td>\n",
       "      <td>0.011827</td>\n",
       "      <td>-0.000892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hispanic</th>\n",
       "      <td>-0.036163</td>\n",
       "      <td>0.023750</td>\n",
       "      <td>-0.066873</td>\n",
       "      <td>-0.015554</td>\n",
       "      <td>-0.025823</td>\n",
       "      <td>-0.005571</td>\n",
       "      <td>0.011167</td>\n",
       "      <td>-0.008214</td>\n",
       "      <td>-0.003607</td>\n",
       "      <td>-0.162269</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.082360</td>\n",
       "      <td>-0.039400</td>\n",
       "      <td>-0.049293</td>\n",
       "      <td>-0.085173</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.057633</td>\n",
       "      <td>-0.555571</td>\n",
       "      <td>-0.012362</td>\n",
       "      <td>0.013951</td>\n",
       "      <td>0.012703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Other_Race</th>\n",
       "      <td>-0.003039</td>\n",
       "      <td>0.012535</td>\n",
       "      <td>0.016205</td>\n",
       "      <td>0.005455</td>\n",
       "      <td>0.005743</td>\n",
       "      <td>0.016102</td>\n",
       "      <td>0.031385</td>\n",
       "      <td>0.008483</td>\n",
       "      <td>0.008762</td>\n",
       "      <td>-0.061319</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.031905</td>\n",
       "      <td>-0.024188</td>\n",
       "      <td>-0.030261</td>\n",
       "      <td>-0.052288</td>\n",
       "      <td>-0.057633</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.341064</td>\n",
       "      <td>-0.013281</td>\n",
       "      <td>0.018433</td>\n",
       "      <td>0.009767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>White</th>\n",
       "      <td>0.040121</td>\n",
       "      <td>-0.049060</td>\n",
       "      <td>0.072367</td>\n",
       "      <td>0.033342</td>\n",
       "      <td>0.001154</td>\n",
       "      <td>-0.003968</td>\n",
       "      <td>-0.024489</td>\n",
       "      <td>-0.015516</td>\n",
       "      <td>0.017314</td>\n",
       "      <td>0.193558</td>\n",
       "      <td>...</td>\n",
       "      <td>0.144862</td>\n",
       "      <td>-0.233164</td>\n",
       "      <td>-0.291707</td>\n",
       "      <td>-0.504040</td>\n",
       "      <td>-0.555571</td>\n",
       "      <td>-0.341064</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.055169</td>\n",
       "      <td>-0.035494</td>\n",
       "      <td>-0.015785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Not_Diabetic</th>\n",
       "      <td>-0.170977</td>\n",
       "      <td>-0.208232</td>\n",
       "      <td>-0.055127</td>\n",
       "      <td>0.057644</td>\n",
       "      <td>-0.102787</td>\n",
       "      <td>-0.153469</td>\n",
       "      <td>-0.032353</td>\n",
       "      <td>-0.209640</td>\n",
       "      <td>0.003636</td>\n",
       "      <td>-0.204576</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.034473</td>\n",
       "      <td>-0.028442</td>\n",
       "      <td>0.002578</td>\n",
       "      <td>-0.055268</td>\n",
       "      <td>-0.012362</td>\n",
       "      <td>-0.013281</td>\n",
       "      <td>0.055169</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.341324</td>\n",
       "      <td>-0.208279</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Borderline_Diabetic</th>\n",
       "      <td>0.016182</td>\n",
       "      <td>0.049231</td>\n",
       "      <td>0.007143</td>\n",
       "      <td>-0.008165</td>\n",
       "      <td>0.011172</td>\n",
       "      <td>0.021437</td>\n",
       "      <td>0.011056</td>\n",
       "      <td>0.031843</td>\n",
       "      <td>-0.004295</td>\n",
       "      <td>0.048657</td>\n",
       "      <td>...</td>\n",
       "      <td>0.006025</td>\n",
       "      <td>0.008356</td>\n",
       "      <td>0.023243</td>\n",
       "      <td>0.011827</td>\n",
       "      <td>0.013951</td>\n",
       "      <td>0.018433</td>\n",
       "      <td>-0.035494</td>\n",
       "      <td>-0.341324</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.013219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Diabetic_during_pregnancy</th>\n",
       "      <td>-0.013930</td>\n",
       "      <td>0.008198</td>\n",
       "      <td>-0.005385</td>\n",
       "      <td>-0.004077</td>\n",
       "      <td>-0.004526</td>\n",
       "      <td>-0.000773</td>\n",
       "      <td>0.018742</td>\n",
       "      <td>-0.007040</td>\n",
       "      <td>-0.085477</td>\n",
       "      <td>-0.033325</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.010092</td>\n",
       "      <td>0.003433</td>\n",
       "      <td>0.007260</td>\n",
       "      <td>-0.000892</td>\n",
       "      <td>0.012703</td>\n",
       "      <td>0.009767</td>\n",
       "      <td>-0.015785</td>\n",
       "      <td>-0.208279</td>\n",
       "      <td>-0.013219</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>25 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                HeartDisease       BMI   Smoking  \\\n",
       "HeartDisease                        1.000000  0.051803  0.107764   \n",
       "BMI                                 0.051803  1.000000  0.023118   \n",
       "Smoking                             0.107764  0.023118  1.000000   \n",
       "AlcoholDrinking                    -0.032080 -0.038816  0.111768   \n",
       "Stroke                              0.196835  0.019733  0.061226   \n",
       "PhysicalHealth                      0.170721  0.109788  0.115352   \n",
       "MentalHealth                        0.028591  0.064131  0.085157   \n",
       "DiffWalking                         0.201258  0.181678  0.120074   \n",
       "Male                                0.070040  0.026940  0.085052   \n",
       "Age                                 0.232325 -0.001740  0.130384   \n",
       "PhysicalActivity                   -0.100030 -0.150616 -0.097174   \n",
       "GenHealth                          -0.243182 -0.230720 -0.174717   \n",
       "SleepTime                           0.008327 -0.051822 -0.030336   \n",
       "Asthma                              0.041444  0.092345  0.024149   \n",
       "KidneyDisease                       0.145197  0.050768  0.034920   \n",
       "SkinCancer                          0.093317 -0.033644  0.033977   \n",
       "American_Indian/Alaskan_Native      0.008547  0.026347  0.035667   \n",
       "Asian                              -0.030262 -0.078643 -0.060308   \n",
       "Black                              -0.010156  0.080672 -0.038249   \n",
       "Hispanic                           -0.036163  0.023750 -0.066873   \n",
       "Other_Race                         -0.003039  0.012535  0.016205   \n",
       "White                               0.040121 -0.049060  0.072367   \n",
       "Not_Diabetic                       -0.170977 -0.208232 -0.055127   \n",
       "Borderline_Diabetic                 0.016182  0.049231  0.007143   \n",
       "Diabetic_during_pregnancy          -0.013930  0.008198 -0.005385   \n",
       "\n",
       "                                AlcoholDrinking    Stroke  PhysicalHealth  \\\n",
       "HeartDisease                          -0.032080  0.196835        0.170721   \n",
       "BMI                                   -0.038816  0.019733        0.109788   \n",
       "Smoking                                0.111768  0.061226        0.115352   \n",
       "AlcoholDrinking                        1.000000 -0.019858       -0.017254   \n",
       "Stroke                                -0.019858  1.000000        0.137014   \n",
       "PhysicalHealth                        -0.017254  0.137014        1.000000   \n",
       "MentalHealth                           0.051282  0.046467        0.287987   \n",
       "DiffWalking                           -0.035328  0.174143        0.428373   \n",
       "Male                                   0.004200 -0.003091       -0.040904   \n",
       "Age                                   -0.058837  0.137280        0.110789   \n",
       "PhysicalActivity                       0.017487 -0.079455       -0.232283   \n",
       "GenHealth                              0.029780 -0.168090       -0.482697   \n",
       "SleepTime                             -0.005065  0.011900       -0.061387   \n",
       "Asthma                                -0.002202  0.038866        0.117907   \n",
       "KidneyDisease                         -0.028280  0.091167        0.142197   \n",
       "SkinCancer                            -0.005702  0.048116        0.041700   \n",
       "American_Indian/Alaskan_Native        -0.004243  0.014357        0.022955   \n",
       "Asian                                 -0.022275 -0.016060       -0.035229   \n",
       "Black                                 -0.025980  0.024819        0.011371   \n",
       "Hispanic                              -0.015554 -0.025823       -0.005571   \n",
       "Other_Race                             0.005455  0.005743        0.016102   \n",
       "White                                  0.033342  0.001154       -0.003968   \n",
       "Not_Diabetic                           0.057644 -0.102787       -0.153469   \n",
       "Borderline_Diabetic                   -0.008165  0.011172        0.021437   \n",
       "Diabetic_during_pregnancy             -0.004077 -0.004526       -0.000773   \n",
       "\n",
       "                                MentalHealth  DiffWalking      Male       Age  \\\n",
       "HeartDisease                        0.028591     0.201258  0.070040  0.232325   \n",
       "BMI                                 0.064131     0.181678  0.026940 -0.001740   \n",
       "Smoking                             0.085157     0.120074  0.085052  0.130384   \n",
       "AlcoholDrinking                     0.051282    -0.035328  0.004200 -0.058837   \n",
       "Stroke                              0.046467     0.174143 -0.003091  0.137280   \n",
       "PhysicalHealth                      0.287987     0.428373 -0.040904  0.110789   \n",
       "MentalHealth                        1.000000     0.152235 -0.100058 -0.155455   \n",
       "DiffWalking                         0.152235     1.000000 -0.068860  0.242552   \n",
       "Male                               -0.100058    -0.068860  1.000000 -0.067682   \n",
       "Age                                -0.155455     0.242552 -0.067682  1.000000   \n",
       "PhysicalActivity                   -0.095808    -0.278524  0.048247 -0.121465   \n",
       "GenHealth                          -0.241625    -0.413797  0.025947 -0.188236   \n",
       "SleepTime                          -0.119717    -0.022216 -0.015704  0.103071   \n",
       "Asthma                              0.114008     0.103222 -0.069191 -0.058235   \n",
       "KidneyDisease                       0.037281     0.153064 -0.009084  0.122697   \n",
       "SkinCancer                         -0.033412     0.064840  0.013434  0.262208   \n",
       "American_Indian/Alaskan_Native      0.018394     0.025205 -0.003434 -0.024922   \n",
       "Asian                              -0.023113    -0.038830  0.013716 -0.084104   \n",
       "Black                               0.010944     0.039615 -0.037280 -0.034563   \n",
       "Hispanic                            0.011167    -0.008214 -0.003607 -0.162269   \n",
       "Other_Race                          0.031385     0.008483  0.008762 -0.061319   \n",
       "White                              -0.024489    -0.015516  0.017314  0.193558   \n",
       "Not_Diabetic                       -0.032353    -0.209640  0.003636 -0.204576   \n",
       "Borderline_Diabetic                 0.011056     0.031843 -0.004295  0.048657   \n",
       "Diabetic_during_pregnancy           0.018742    -0.007040 -0.085477 -0.033325   \n",
       "\n",
       "                                ...  SkinCancer  \\\n",
       "HeartDisease                    ...    0.093317   \n",
       "BMI                             ...   -0.033644   \n",
       "Smoking                         ...    0.033977   \n",
       "AlcoholDrinking                 ...   -0.005702   \n",
       "Stroke                          ...    0.048116   \n",
       "PhysicalHealth                  ...    0.041700   \n",
       "MentalHealth                    ...   -0.033412   \n",
       "DiffWalking                     ...    0.064840   \n",
       "Male                            ...    0.013434   \n",
       "Age                             ...    0.262208   \n",
       "PhysicalActivity                ...   -0.001328   \n",
       "GenHealth                       ...   -0.050641   \n",
       "SleepTime                       ...    0.041266   \n",
       "Asthma                          ...   -0.000396   \n",
       "KidneyDisease                   ...    0.061816   \n",
       "SkinCancer                      ...    1.000000   \n",
       "American_Indian/Alaskan_Native  ...   -0.026784   \n",
       "Asian                           ...   -0.047749   \n",
       "Black                           ...   -0.083390   \n",
       "Hispanic                        ...   -0.082360   \n",
       "Other_Race                      ...   -0.031905   \n",
       "White                           ...    0.144862   \n",
       "Not_Diabetic                    ...   -0.034473   \n",
       "Borderline_Diabetic             ...    0.006025   \n",
       "Diabetic_during_pregnancy       ...   -0.010092   \n",
       "\n",
       "                                American_Indian/Alaskan_Native     Asian  \\\n",
       "HeartDisease                                          0.008547 -0.030262   \n",
       "BMI                                                   0.026347 -0.078643   \n",
       "Smoking                                               0.035667 -0.060308   \n",
       "AlcoholDrinking                                      -0.004243 -0.022275   \n",
       "Stroke                                                0.014357 -0.016060   \n",
       "PhysicalHealth                                        0.022955 -0.035229   \n",
       "MentalHealth                                          0.018394 -0.023113   \n",
       "DiffWalking                                           0.025205 -0.038830   \n",
       "Male                                                 -0.003434  0.013716   \n",
       "Age                                                  -0.024922 -0.084104   \n",
       "PhysicalActivity                                     -0.017736  0.012106   \n",
       "GenHealth                                            -0.037600  0.032701   \n",
       "SleepTime                                            -0.003615 -0.019985   \n",
       "Asthma                                                0.013757 -0.017007   \n",
       "KidneyDisease                                         0.007401 -0.016957   \n",
       "SkinCancer                                           -0.026784 -0.047749   \n",
       "American_Indian/Alaskan_Native                        1.000000 -0.020687   \n",
       "Asian                                                -0.020687  1.000000   \n",
       "Black                                                -0.035746 -0.044721   \n",
       "Hispanic                                             -0.039400 -0.049293   \n",
       "Other_Race                                           -0.024188 -0.030261   \n",
       "White                                                -0.233164 -0.291707   \n",
       "Not_Diabetic                                         -0.028442  0.002578   \n",
       "Borderline_Diabetic                                   0.008356  0.023243   \n",
       "Diabetic_during_pregnancy                             0.003433  0.007260   \n",
       "\n",
       "                                   Black  Hispanic  Other_Race     White  \\\n",
       "HeartDisease                   -0.010156 -0.036163   -0.003039  0.040121   \n",
       "BMI                             0.080672  0.023750    0.012535 -0.049060   \n",
       "Smoking                        -0.038249 -0.066873    0.016205  0.072367   \n",
       "AlcoholDrinking                -0.025980 -0.015554    0.005455  0.033342   \n",
       "Stroke                          0.024819 -0.025823    0.005743  0.001154   \n",
       "PhysicalHealth                  0.011371 -0.005571    0.016102 -0.003968   \n",
       "MentalHealth                    0.010944  0.011167    0.031385 -0.024489   \n",
       "DiffWalking                     0.039615 -0.008214    0.008483 -0.015516   \n",
       "Male                           -0.037280 -0.003607    0.008762  0.017314   \n",
       "Age                            -0.034563 -0.162269   -0.061319  0.193558   \n",
       "PhysicalActivity               -0.043701 -0.056564    0.001808  0.064172   \n",
       "GenHealth                      -0.062215 -0.028721   -0.015241  0.062656   \n",
       "SleepTime                      -0.020192 -0.013242   -0.031007  0.042905   \n",
       "Asthma                          0.021297  0.003885    0.028037 -0.025421   \n",
       "KidneyDisease                   0.010300 -0.010960   -0.002057  0.005933   \n",
       "SkinCancer                     -0.083390 -0.082360   -0.031905  0.144862   \n",
       "American_Indian/Alaskan_Native -0.035746 -0.039400   -0.024188 -0.233164   \n",
       "Asian                          -0.044721 -0.049293   -0.030261 -0.291707   \n",
       "Black                           1.000000 -0.085173   -0.052288 -0.504040   \n",
       "Hispanic                       -0.085173  1.000000   -0.057633 -0.555571   \n",
       "Other_Race                     -0.052288 -0.057633    1.000000 -0.341064   \n",
       "White                          -0.504040 -0.555571   -0.341064  1.000000   \n",
       "Not_Diabetic                   -0.055268 -0.012362   -0.013281  0.055169   \n",
       "Borderline_Diabetic             0.011827  0.013951    0.018433 -0.035494   \n",
       "Diabetic_during_pregnancy      -0.000892  0.012703    0.009767 -0.015785   \n",
       "\n",
       "                                Not_Diabetic  Borderline_Diabetic  \\\n",
       "HeartDisease                       -0.170977             0.016182   \n",
       "BMI                                -0.208232             0.049231   \n",
       "Smoking                            -0.055127             0.007143   \n",
       "AlcoholDrinking                     0.057644            -0.008165   \n",
       "Stroke                             -0.102787             0.011172   \n",
       "PhysicalHealth                     -0.153469             0.021437   \n",
       "MentalHealth                       -0.032353             0.011056   \n",
       "DiffWalking                        -0.209640             0.031843   \n",
       "Male                                0.003636            -0.004295   \n",
       "Age                                -0.204576             0.048657   \n",
       "PhysicalActivity                    0.135451            -0.017253   \n",
       "GenHealth                           0.273381            -0.047783   \n",
       "SleepTime                          -0.000494            -0.006957   \n",
       "Asthma                             -0.050782             0.016750   \n",
       "KidneyDisease                      -0.142371             0.004982   \n",
       "SkinCancer                         -0.034473             0.006025   \n",
       "American_Indian/Alaskan_Native     -0.028442             0.008356   \n",
       "Asian                               0.002578             0.023243   \n",
       "Black                              -0.055268             0.011827   \n",
       "Hispanic                           -0.012362             0.013951   \n",
       "Other_Race                         -0.013281             0.018433   \n",
       "White                               0.055169            -0.035494   \n",
       "Not_Diabetic                        1.000000            -0.341324   \n",
       "Borderline_Diabetic                -0.341324             1.000000   \n",
       "Diabetic_during_pregnancy          -0.208279            -0.013219   \n",
       "\n",
       "                                Diabetic_during_pregnancy  \n",
       "HeartDisease                                    -0.013930  \n",
       "BMI                                              0.008198  \n",
       "Smoking                                         -0.005385  \n",
       "AlcoholDrinking                                 -0.004077  \n",
       "Stroke                                          -0.004526  \n",
       "PhysicalHealth                                  -0.000773  \n",
       "MentalHealth                                     0.018742  \n",
       "DiffWalking                                     -0.007040  \n",
       "Male                                            -0.085477  \n",
       "Age                                             -0.033325  \n",
       "PhysicalActivity                                 0.001501  \n",
       "GenHealth                                       -0.001201  \n",
       "SleepTime                                       -0.011668  \n",
       "Asthma                                           0.013282  \n",
       "KidneyDisease                                   -0.004706  \n",
       "SkinCancer                                      -0.010092  \n",
       "American_Indian/Alaskan_Native                   0.003433  \n",
       "Asian                                            0.007260  \n",
       "Black                                           -0.000892  \n",
       "Hispanic                                         0.012703  \n",
       "Other_Race                                       0.009767  \n",
       "White                                           -0.015785  \n",
       "Not_Diabetic                                    -0.208279  \n",
       "Borderline_Diabetic                             -0.013219  \n",
       "Diabetic_during_pregnancy                        1.000000  \n",
       "\n",
       "[25 rows x 25 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Look at the correlation of the data\n",
    "df.corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking at the correlation of each variable with Heart Disease, seems like there isn't a specific variable that highly correlat with heart disease.\n",
    "\n",
    "Below, we decided to visualize the correlation of smoking with heart disease and the correlation of age vs heart disease. These two are among the common features that causes heart diseases."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Percentage of Heart Disease Patients that Smoke vs Does not Smoke')"
      ]
     },
     "execution_count": 10,
     "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": [
    "fig, ax = plt.subplots(figsize = (6,4))\n",
    "\n",
    "heart_disease_smoking_cnt = len(df[(df['HeartDisease']==1) & (df['Smoking']==1)]['Smoking'])\n",
    "smoking_cnt = len(df[df['Smoking']==1]['Smoking'])\n",
    "heart_disease_no_smoking_cnt = len(df[(df['HeartDisease']==1) & (df['Smoking']==0)]['Smoking'])\n",
    "non_smoking_cnt = len(df[df['Smoking']==0]['Smoking'])\n",
    "\n",
    "smoking_heart_diseases = heart_disease_smoking_cnt/smoking_cnt\n",
    "not_smoking_heart_diseases = heart_disease_no_smoking_cnt/non_smoking_cnt\n",
    "\n",
    "ax.bar([\"Smoking\", \"Non-smoking\"], [smoking_heart_diseases, not_smoking_heart_diseases], color=['r','b'])\n",
    "ax.set_xlabel(\"patient smoking status\")\n",
    "ax.set_ylabel(\"Proportion of Heart disease\")\n",
    "plt.title(\"Percentage of Heart Disease Patients that Smoke vs Does not Smoke\")"
   ]
  },
  {
   "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": [
    "# visualize the proportion of heart disease vs age\n",
    "x_label = ['18', '25', '30', '35', '40', '45', '50', '55', '60', '65', '70', '75', '80']\n",
    "\n",
    "count_by_age_df = df[\"Age\"].value_counts()\n",
    "count_by_age_arr = [count_by_age_df[int(i)] for i in x_label]\n",
    "\n",
    "count_by_hd_age_df = df[df[\"HeartDisease\"]==1][\"Age\"].value_counts()\n",
    "count_by_hd_age_arr = [count_by_hd_age_df[int(i)] for i in x_label]\n",
    "\n",
    "proportion = [count_by_hd_age_arr[i]/count_by_age_arr[i] for i in range(len(x_label))]\n",
    "\n",
    "plt.figure(figsize=(6, 4))\n",
    "plt.bar(x_label, proportion)\n",
    "plt.xlabel(\"Age Groups\")\n",
    "plt.ylabel(\"% of Heart Disease Instances\")\n",
    "plt.title(\"Proportion of Heart Disease patients by Age\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Proposed Solution\n",
    "\n",
    "We will be developing a logistic regression, K-NN, Neuro Network, and Random Forest model, given parameters, such as smoking, alcohol drinking, and etc., to predict whether a patient is likely to have heart diseases or not. The goal of having these models is to find the best model that can make heart disease predictions. With logistic regression, we can also make inference on which variable influences the decision.\n",
    "\n",
    "By discovering similar characteristics for patients who have heart disease, along with a substantial number of observations, we hope to apply some of these characteristics as a more general prediction for patients who could have heart disease in the future as a preventative measure."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluation Metrics\n",
    "\n",
    "10% out of 320,000 observations in the total dataset will be used toward the test set. Note the proportion may be changed depending on how we modify our approach (we may need to reduce the number of patients with no heart disease to balance out the heart disease variable proportion).\n",
    "\n",
    "The F1 score best fits our research question because it measures how correctly our data is classified without putting a heavier weight on True Positives or False Negatives since it is a harmonic mean between precision and recall. Additionally, we chose not to use the accuracy metric because our data is pretty unbalanced, with a very low proportion of patients with heart disease which can lead to a biased or misleading interpretation of results when using the accuracy metric. Therefore, F1 provides a more holistic interpretation of results, as it takes into account both precision and recall, while not be as affected by our unbalanced data like accuracy would. However, recall score being much more important than percision will also be monitored throughout the project."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Additional Imports\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Split\n",
    "\n",
    "Split the data into train/test, we will use stratified shuffle split as people who have heart disease heavily outweights people who does not."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((287815, 24), (31980, 24), (287815,), (31980,))"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# Note Column Names Are Not Included\n",
    "data = df.to_numpy()\n",
    "X = data[:,1:]\n",
    "y = data[:,0]\n",
    "\n",
    "# Split data (with stratified shuffle split)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.10, random_state=1, stratify=y)\n",
    "\n",
    "# Output shape of each dataset\n",
    "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Baseline Preformance Estimation\n",
    "#### Logistic Regression 1.0\n",
    "\n",
    "Here, we will preform a simple logistic regression to provide a baseline of what the most general case of logistic regression can provide."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "logreg = LogisticRegression(max_iter=5000, random_state=1)\n",
    "clf = logreg.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.92      0.99      0.96     29243\n",
      "         1.0       0.56      0.11      0.19      2737\n",
      "\n",
      "    accuracy                           0.92     31980\n",
      "   macro avg       0.74      0.55      0.57     31980\n",
      "weighted avg       0.89      0.92      0.89     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_test_pred = clf.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As shown above, the test-accuracy seems very good with a score of 0.92. However, if we look closely, the recall and f1-score is terrible being 0.11 and 0.19."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our baseline performance preforms extremely poorly. As shown by the confusion matrix above, logistic regression works decently well when looking at accuracy. However, the classification of whether a patient may have heart disease is more important when the patient actually has heart disease. When looking at its precision and recall, we can see that logistic regression does a poor job of classifying those with heart disease properly. Logistic regression only classified 56% of patients as having heart disease correctly out of all patients that it had classified as having heart disease (precision). Also, only a measly 11% were properly classified as having heart disease out of all patients who did (recall). With our focus on classification of patients having heart disease, recall is a very important metric as those with heart disease should be properly identified as having heart disease. In addition, the f1 score obtained is 0.19, this further shows our baseline algorithm's inability to predict correctly."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Logistic Regression Continued: Attempt to fix the issue\n",
    "\n",
    "To fix the issue of unbalanced dataset in terms of labeling, there are few solutions on our mind. Below, we will be trying one of them by introducing class weight to logistic regression:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.97      0.75      0.85     29243\n",
      "         1.0       0.23      0.78      0.35      2737\n",
      "\n",
      "    accuracy                           0.75     31980\n",
      "   macro avg       0.60      0.76      0.60     31980\n",
      "weighted avg       0.91      0.75      0.80     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "logreg = LogisticRegression(max_iter=5000, random_state=1, class_weight=\"balanced\")\n",
    "clf = logreg.fit(X_train, y_train)\n",
    "\n",
    "y_test_pred = clf.predict(X_test)\n",
    "\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Above, we introduced class_weight with the \"balanced\" setting that sklearn have. Without dividng deep into the dataset (for now), this setting give us a much better recall value, while prevision worsten. However, we view this as an significant improvement as the more important recall score went from 0.11 (without class_weight) to 0.78 (with class_weight). F1 score also have an improvement through the combination of the significant increment of recall and decrease of precision. F1 score went from 0.19 (without class_weight) to 0.35 (with class_weight).\n",
    "\n",
    "Note: there are still a significant room for improvement. There are more things we can try such as remove some of the healthy patient data to achieve balance, try out different algorithem, and manually adjust class weight."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### KNN 1.0\n",
    "\n",
    "We also tried using KNN and see if the result would improve, however both f1 and recall scores preformed poorly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "num_neigh = [*range(1, 10, 2)]\n",
    "recall_scores = []\n",
    "f1_scores = []\n",
    "\n",
    "for num in num_neigh:\n",
    "    neigh = KNeighborsClassifier(n_neighbors = num)\n",
    "    neigh.fit(X_train, y_train)\n",
    "    y_pred = neigh.predict(X_test)\n",
    "    recall_score = metrics.recall_score(y_test, y_pred)\n",
    "    f1_score = metrics.f1_score(y_test, y_pred)\n",
    "    recall_scores.append(recall_score)\n",
    "    f1_scores.append(f1_score)\n",
    "# want to minimize false negative and maximize true positive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x2d081ebd850>"
      ]
     },
     "execution_count": 18,
     "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": [
    "plt.plot(num_neigh, recall_scores, label = 'Recall score')\n",
    "plt.plot(num_neigh, f1_scores, label = 'f1 score')\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel('Number of Neighbors')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Neuro Network 1.0\n",
    "As shown below, neuro network model completely fails to identify patients with heart disease. We obtained an f1 score of 0.00 as Neuro Network decided to classify every patient as having no heart disease."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.91      1.00      0.96     29243\n",
      "         1.0       0.00      0.00      0.00      2737\n",
      "\n",
      "    accuracy                           0.91     31980\n",
      "   macro avg       0.46      0.50      0.48     31980\n",
      "weighted avg       0.84      0.91      0.87     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neural_network import MLPClassifier\n",
    "nn = MLPClassifier(solver='sgd', alpha=1e-5,\n",
    "                    hidden_layer_sizes=(5, 3), random_state=1, max_iter=1000)\n",
    "nn.fit(X_train, y_train)\n",
    "y_test_pred = nn.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred, zero_division=0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Random Forest 1.0\n",
    "This algorithem also did not preform well (even with balance setting on class_weight!)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.92      0.98      0.95     29243\n",
      "         1.0       0.31      0.12      0.17      2737\n",
      "\n",
      "    accuracy                           0.90     31980\n",
      "   macro avg       0.62      0.55      0.56     31980\n",
      "weighted avg       0.87      0.90      0.88     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rf = RandomForestClassifier(criterion='gini',n_estimators=25, class_weight=\"balanced\")\n",
    "\n",
    "rf.fit(X_train, y_train)\n",
    "y_test_pred = rf.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Interesting thing to note is having class_weight=\"balanced\" for logistic regression improved the recall and f1-score significantly, but this is not the case for RF.\n",
    "\n",
    "Seems like the problem of unbalanced data occurs throughout all the algorithems presented. We would need to fix the unbalanced data issue."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Re-handling\n",
    "\n",
    "Seeing the poor results from the algorithms above, we decided to balance out the data by preofrming Oversampling and Standard Scaler."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Additional imports\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from imblearn.over_sampling  import RandomOverSampler\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Obtain X and Y dataframe\n",
    "X = df.drop('HeartDisease',axis=1)\n",
    "y = df['HeartDisease']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We want to split the data first, then preform oversampling to avoid artificially inflate the result score."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split data\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y,test_size=0.10,shuffle=True,random_state=1, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    263179\n",
       "1     24636\n",
       "Name: HeartDisease, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preform Oversampling\n",
    "\n",
    "As seen above, our label's positive (1) case is very under-represented. We will attempt to balance the data by random oversampling. By random-oversampling, we balanced the variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    263179\n",
       "1    263179\n",
       "Name: HeartDisease, dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sampler = RandomOverSampler(random_state=1)\n",
    "X_train, y_train = sampler.fit_resample(X_train,y_train)\n",
    "\n",
    "y_train.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preform Data Standardization\n",
    "\n",
    "Looking at the histogram at EDA, we can see that many features are quite unbalanced. For example, there are a lot more people who don't have kidney disease than people who does. By standardizing the features weights will now be standardized."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "standardizer = StandardScaler()\n",
    "X_train = standardizer.fit_transform(X_train)\n",
    "\n",
    "# transform the test set as well\n",
    "X_test = standardizer.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Algorithm Preformances\n",
    "#### Logistic Regression 2.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.97      0.75      0.85     29243\n",
      "           1       0.23      0.78      0.35      2737\n",
      "\n",
      "    accuracy                           0.75     31980\n",
      "   macro avg       0.60      0.76      0.60     31980\n",
      "weighted avg       0.91      0.75      0.80     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Logistic Regression 2.0, with updated data\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "logreg = LogisticRegression(max_iter=5000, random_state=1)\n",
    "clf = logreg.fit(X_train, y_train)\n",
    "\n",
    "y_test_pred = clf.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Logistic Regression have the same result as above when weight_balance is set to 'balanced', this make sense as data standardization is preforming the same thing (adjusting weights).\n",
    "\n",
    "#### K-NN 2.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# KNN 2.0, with updated data\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "num_neigh = [*range(1, 20, 2)]\n",
    "recall_scores = []\n",
    "f1_scores = []\n",
    "\n",
    "for num in num_neigh:\n",
    "    neigh = KNeighborsClassifier(n_neighbors = num)\n",
    "    neigh.fit(X_train, y_train)\n",
    "    y_pred = neigh.predict(X_test)\n",
    "    recall_score = metrics.recall_score(y_test, y_pred)\n",
    "    f1_score = metrics.f1_score(y_test, y_pred)\n",
    "    recall_scores.append(recall_score)\n",
    "    f1_scores.append(f1_score)\n",
    "# want to minimize false negative and maximize true positive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max recall score: 0.7175739861161856\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAAxwUlEQVR4nO3dd3hUZdrH8e+dRgiQhN5CCRh6J4liAVSaZQXWRQRUsCEW1H1XVlxfXffdYlsbiCAqoljABjaaIHZa6C2EDkkEQktISJ153j/OBIc4IQPM5CSZ+3Ndc82cc55z5s5hmN+c9hwxxqCUUipwBdldgFJKKXtpECilVIDTIFBKqQCnQaCUUgFOg0AppQJciN0FnKt69eqZli1b2l2GUkpVKmvWrDlijKnvaVqlC4KWLVuSlJRkdxlKKVWpiMi+0qbpriGllApwGgRKKRXgNAiUUirAVbpjBJ4UFhaSmppKXl6e3aUEnPDwcGJiYggNDbW7FKXUeaoSQZCamkqtWrVo2bIlImJ3OQHDGMPRo0dJTU0lNjbW7nKUUuepSuwaysvLo27duhoC5UxEqFu3rm6JKVXJVYkgADQEbKLrXanKr0rsGlJKKW85nYbcQof1KPjt+VSBgzzX+FOnxxeRW+DEYQwhQUJw8UOEoCAhJOi352Bxm16ibXCw9ezePqjEPGdMEyEk+PfLrBYSTFiI73+/axD4SHBwMJ07d6aoqIjY2FhmzZpFdHS0z5ZffCFdvXr1qFmzJtnZ2T5btlIVUW6Bg/TM3DO+rN2fT39xF5T44i50kFvoPP36VIGDPLd58oucdv9p5+2ePq147Jr2Pl+uBoGPVK9enfXr1wMwevRopkyZwuOPP25vURegqKiIkBD9eKjykZlbyNb0LLakZ7I5LZMt6VnsysjG6cV9s4IEIsJCqB4WTPVQ18P1umGtUMJdryPCzpzm/hwRFkx4aHG7kN+mhwUTHhJEcJDgNFDkdOJ0gsMYHA6Dw5jT4854NoYip8FR2sM13VliXPHr4mlFTmMty2E9d24a5Zf1r//T/aBXr15s3LgRgF27dnH//feTkZFBREQEb7zxBu3atePQoUOMGzeO3bt3AzB16lQuvfRShgwZwoEDB8jLy+Ohhx5i7NixXr1nTk4ON910E6mpqTgcDp544gmGDx/O6tWreeihh8jJyaFatWosXbqU0NBQ7r33XpKSkggJCeHFF1/kyiuvZObMmXz99dfk5eWRk5PDl19+yfjx49m0aRNFRUU89dRTDB482G/rTQWGjJP5bE7PZGt61ukv/f3HTp2e3igynI5NIrmmc2Ni60UQERbi8Us8IjSE8LAgwoKDyuVYVbBAcFCw39/HDlUuCP7x5Ra2pmf5dJkdmkTy9z909Kqtw+Fg6dKl3HnnnQCMHTuWadOmERcXx8qVK7nvvvv49ttvefDBB+nTpw9z587F4XCc3tUzY8YM6tSpQ25uLgkJCdx4443UrVu3zPdduHAhTZo04euvvwYgMzOTgoIChg8fzpw5c0hISCArK4vq1avzyiuvALBp0yaSk5MZMGAAKSkpACxfvpyNGzdSp04d/va3v3HVVVcxY8YMTpw4QWJiIv369aNGjRrnvA5V4DHGkHYil81pWWxNz2Sz6xf/oaz8021a1I2gU9NIhic0o1PTKDo2iaRezWo2Vh2Y/BoEIjIIeAUIBt40xjxTYvoEYJRbLe2B+saYY/6syx9yc3Pp1q0be/fupWfPnvTv35/s7Gx++eUXhg0bdrpdfr71n+Dbb7/l3XffBazjC1FR1ibfpEmTmDt3LgAHDhxgx44dXgVB586deeSRR3j00Ue5/vrrueKKK9i0aRONGzcmISEBgMjISAB++uknxo8fD0C7du1o0aLF6SDo378/derUAWDx4sV88cUX/Pe//wWs03T3799P+/a+30epKjen07DnaA6b01y/9NOtX/onThUC1u6bixrU5LLW9ejQJJJOTaPo0CSSyHC9ELEi8FsQiEgwMAXoD6QCq0XkC2PM1uI2xpjngedd7f8A/PlCQ8DbX+6+VnyMIDMzk+uvv54pU6YwZswYoqOjTx87KMt3333HkiVLWL58OREREfTt29frc/TbtGnDmjVrmD9/Po899hgDBgxgyJAhHjeZjSl9x6v7r31jDJ9++ilt27b1qgYVGAodTnYcyj5j9862X7PIKXAAEBYcRNtGtbimUyM6NImiU5NI2jWKpHpY1dytUhX4c4sgEdhpjNkNICKzgcHA1lLajwA+9GM95SIqKopJkyYxePBg7r33XmJjY/n4448ZNmwYxhg2btxI165dufrqq5k6dSoPP/wwDoeDnJwcMjMzqV27NhERESQnJ7NixQqv3zc9PZ06depwyy23ULNmTWbOnMnEiRNJT09n9erVJCQkcPLkSapXr07v3r15//33ueqqq0hJSWH//v20bduWtWvXnrHMgQMHMnnyZCZPnoyIsG7dOrp37+7rVaYqsNwCB8kHs9ic7tq9k5bF9oMnKXBYZ95EhAXToXEkw+KbWb/0m0QR17AmocFV5hKlgODPIGgKHHAbTgUu9tRQRCKAQcADfqyn3HTv3p2uXbsye/Zs3n//fe69917+9a9/UVhYyM0330zXrl155ZVXGDt2LG+99RbBwcFMnTqVQYMGMW3aNLp06ULbtm255JJLvH7PTZs2MWHCBIKCgggNDWXq1KmEhYUxZ84cxo8fT25uLtWrV2fJkiXcd999jBs3js6dOxMSEsLMmTOpVu33+2WfeOIJHn74Ybp06YIxhpYtW/LVV1/5clWpCiav0MGSbYf4dtthNqdnsvPwb2fuREeE0rFJJLdf1pKOrv35LevWIDhILyqs7ORsuwkuaMEiw4CBxpi7XMO3AonGmPEe2g4HbjHG/KGUZY0FxgI0b9685759Z95fYdu2bbrf2ka6/is3p9OwYs9R5q1LY8Gmg5zML6JezTC6xkTTsUnk6S/9ptHV9UrySkxE1hhj4j1N8+cWQSrQzG04Bkgvpe3NnGW3kDFmOjAdID4+3j/JpVSAST6Yxdx1aXyxPp1fM/OoWS2Eazo1Ymj3plzcqq7+0g8g/gyC1UCciMQCaVhf9iNLNhKRKKAPcIsfa1FKAQcz8/hiQxpz16Wz7dcsQoKEPm3q87dr29OvfUM9oBug/BYExpgiEXkAWIR1+ugMY8wWERnnmj7N1XQosNgYk+OvWpQKZNn5RSzcfJB569L4edcRjIFuzaL5xw0dub5LY+rqefsBz6/XERhj5gPzS4ybVmJ4JjDTn3UoFWgKHU5+3JHB3HXpfLP1IHmFTlrUjeDBq+IY0r0psfX0okD1myp3ZbFSgcoYw4bUTOatS+PLDekczSkgOiKUYT2bMaR7U3o0j9aDvcojDQKlKrn9R08xd10a89ansedIDmEhQfRv35Ah3ZvSp019v3RbrKoW/YT4yKRJk2jfvj2jRo0iOTmZXr16Ua1atdPdMyjlS8dzCpi1Yh83Tv2F3s8v46UlKTSMrMazN3Ym6X/7MWVUD/p3aKghoLyiWwQ+8tprr7FgwQJiY2M5fPgwkyZNYt68eeX2/g6Hg+BgPeOjKssrdPBt8mE+W5vG9ymHKXQY2jSsyaOD2nFDtyY0ja5ud4mqktKfCz5Q3J30DTfcwEsvvUSDBg1ISEggNLT0DrUcDgdjxoyhU6dOdO7cmZdeegmAnTt30q9fP7p27UqPHj3YtWsXxhgmTJhwuu2cOXMAq2+iK6+8kpEjR9K5c2ccDgcTJkwgISGBLl268Prrr5fL36/8x+k0LN91lEc/2UjCv5dw3/tr2Zh6gjGXtuTrBy9n0cO9ubdvaw0BdUGq3hbBgolwcJNvl9moM1zzTKmTp02bxsKFC1m2bBn16tXzapHr168nLS2NzZs3A3DixAkARo0axcSJExk6dCh5eXk4nU4+++wz1q9fz4YNGzhy5AgJCQn07t0bgFWrVrF582ZiY2OZPn06UVFRrF69mvz8fC677DIGDBhAbGzshf39qtylHDrJ3HVpfL4ujfTMPCLCghnkutjr0tb19GIv5VNVLwgqiVatWrF7927Gjx/Pddddx4ABAzh58iRpaWkMHToUgPDwcMDqNnrEiBEEBwfTsGFD+vTpw+rVq4mMjCQxMfH0F/3ixYvZuHEjn3zyCWDdk2DHjh0aBJVEdn4Rs1ft57O1aWz9NYvgIOGKuHo8ek07+ndoSESY/ndV/lH1Plln+eVekdSuXZsNGzawaNEipkyZwkcffcTLL7/sse25dBs9efJkBg4c6OtylR8VFDmZvXo/k5bu4Eh2AV1jovj7HzpwfZcm1K+lF3sp/9NjBDY5cuQITqeTG2+8kX/+85+sXbuWyMhIYmJiTh9kzs/P59SpU/Tu3Zs5c+bgcDjIyMjghx9+IDEx8XfLHDhwIFOnTqWw0LoZSEpKCjk5esF2ReV0Gr7ckE7/l77nyc+30Lp+TebedymfP3A5t18WqyGgyk3V2yKoAA4ePEh8fDxZWVkEBQXx8ssvs3Xr1tN3CANIS0vj9ttvx+m0+nV/+umnAZg1axb33HMPTz75JKGhoXz88ccMHTqU5cuX07VrV0SE5557jkaNGpGcnHzG+951113s3buXHj16YIyhfv365XrmkvLeLzuP8MzCZDamZtK2YS3eHpNA37b19YIvZQu/dUPtL/Hx8SYpKemMcdoNsr10/Xtva3oWzy5M5vuUDJpEhfM/A9oytHtTPfir/M6ubqiVUi4Hjp3ixW9SmLc+jcjwUB6/tj239mpBeKhe+6Hsp0GglB8dzyng1WU7mbV8HyJwT+/W3NunNVERetN2VXFUmSAwxuj+VRtUtl2L5SW3wMGMn/cw7btd5BQU8aeeMTzcrw1N9MIvVQFViSAIDw/n6NGj1K1bV8OgHBljOHr06OnrHRQUOZx8siaVl5akcCgrn37tG/LXQW1p07CW3aUpVaoqEQQxMTGkpqaSkZFhdykBJzw8nJiYGLvLsJ0xhm+2HuK5RdvZeTibHs2jmTyiB4mxdewuTakyVYkgCA0N1atnlW2S9h7j6QXJrNl3nFb1azDtlp4M7NhQt05VpVElgkApO+w4dJLnFm3nm62HaFCrGk//sTPDesYQEqzXaarKRYNAqXN0MDOPl75J4eM1B6gRFsKEgW25/bKW2heQqrT0k6uUlzJzC5n2/S5m/LQHpzGMuTSWB666iDo1wuwuTakLokGgVBnyCh28t2Ifry7bSWZuIUO6NeV/+rehWZ0Iu0tTyic0CJQqhcNp+Hx9Gi8sTiHtRC6929Tn0UFt6dgkyu7SlPIpDQKlSjDG8H1KBs8sSCb54Ek6N43iuT914bKLvLvpkFKVjQaBUm42HDjBMwuSWb77KM3rRDB5RHeu69yYIO0UTlVhGgRKAXuP5PD84u18vfFX6tYI4x83dGREYnPCQvRUUFX1aRCogJZf5GDKsl1M/W4nocFBPHh1HGN7t6JmNf2voQKHftpVwFq15xgTP9vI7owchnZvymPXtqNBLe03SQUeDQIVcDJzC3lmQTIfrtpPTO3qvHNHIn3a1Le7LKVso0GgAoYxhgWbD/L3L7ZwNDufsb1b8XC/OL0iWAU8/R+gAkL6iVye/HwLS7YdolPTSN4ek0Cnpno9gFKgQaCqOIfT8N6KfTy3MBmngcevbc/tl7XUjuGUcqNBoKqs5INZTPx0E+sPnKB3m/r8e0gn7RZCKQ80CFSVk1foYPK3O3j9+91EVg/l5eHdGNytid4fQKlS+DUIRGQQ8AoQDLxpjHnGQ5u+wMtAKHDEGNPHnzWpqu2XXUd4fO5m9hzJ4cYeMfzvde2prb2DKnVWfgsCEQkGpgD9gVRgtYh8YYzZ6tYmGngNGGSM2S8iDfxVj6raTpwq4D/zt/FRUirN60Tw3p0Xc3mc9g2klDf8uUWQCOw0xuwGEJHZwGBgq1ubkcBnxpj9AMaYw36sR1VBxhi+3Pgr//flFo6fKmRcn9Y8dHUc1cOC7S5NqUrDn0HQFDjgNpwKXFyiTRsgVES+A2oBrxhj3i25IBEZC4wFaN68uV+KVZVP6vFTPDFvM8u2Z9A1Jop377iYDk0i7S5LqUrHn0Hg6cic8fD+PYGrgerAchFZYYxJOWMmY6YD0wHi4+NLLkMFGIfTMPOXvbyweDsAT1zfgTGXtiRYewhV6rz4MwhSgWZuwzFAuoc2R4wxOUCOiPwAdAVSUMqDLemZPPbZJjamZnJl2/r8c0gnYmrrKaFKXQh/BsFqIE5EYoE04GasYwLuPgdeFZEQIAxr19FLfqxJVVK5BQ5eXprCmz/uoXZEKJNHdOf6Lo31lFClfMBvQWCMKRKRB4BFWKePzjDGbBGRca7p04wx20RkIbARcGKdYrrZXzWpyumnHUf429xN7D92ipviY/jbte2JjtBTQpXyFTGmcu1yj4+PN0lJSXaXocrBsZwC/vX1Vj5bm0ZsvRr8e2gnLm2tp4QqdT5EZI0xJt7TNL2yWFU4xhjmrU/jn19tIyu3kPuvbM34q+IID9VTQpXyBw0CVaEcOHaKx+dt5oeUDLo1i+aZGzvTrpGeEqqUP2kQqAqhyOFkxs97ePGbFIJF+McNHbnlkhZ6SqhS5UCDQNluc1omj366kS3pWfRr34D/G9yJJtHV7S5LqYChQaBs43QaXvwmhde+20ndmtV4bVQPrunUSE8JVaqcaRAoWzidhsfnbebDVfv5U88YnriuA1ERoXaXpVRA0iBQ5c7hNDz66UY+WZPKfX1bM2FgW90KUMpGGgSqXBU5nDzy8QbmrU/noavjeLhfnIaAUjbTIFDlptDh5M9z1vPVxl95ZEAbHrgqzu6SlFJoEKhyUlDk5KHZ61iw+SCPXdOOe/q0trskpZSLBoHyu/wiB/e/v44l2w7x5PUduOPyWLtLUkq50SBQfpVX6GDce2v4bnsG/xzckVt7tbS7JKVUCRoEym9yCxyMnZXETzuP8PQfOzMiUe8up1RFpEGg/CInv4g731nNyj3HeO7GLgyLb1b2TEopW2gQKJ/Lzi/i9rdXsWbfcV66qRtDuje1uySl1FloECifysorZPSMVWxMzWTSiO5c36WJ3SUppcqgQaB8JvNUIbfOWMm2X7OYMrIHgzo1srskpZQXNAiUTxzPKeCWt1ay41A2U0f1pF+HhnaXpJTykgaBumBHsvO55c2V7D6Sw/TbetK3bQO7S1JKnQMNAnVBDp/MY9QbKzlw/BQzRidweZzeU1ipykaDQJ23g5l5jHxjBQez8ph5eyKXtKprd0lKqfOgQaDOS9qJXEa+sYIjJ/N5545EElrWsbskpdR50iBQ5+zAsVOMeGMFmacKmXXXxfRoXtvukpRSF0CDQJ2TfUdzGDF9Bdn5Rbx/98V0iYm2uySl1AXSIFBe252Rzcg3VpJf5OCDuy+hU9Mou0tSSvmABoHyys7DJxnxxkqcTsOHYy+hXaNIu0tSSvlIkN0FqIov+WAWw19fAcBsDQGlqhzdIlBntSU9k1veXEm1kGA+uPtiWtWvaXdJSikf0yBQpdqYeoJb31pFzWohfHD3xbSoW8PukpRSfqC7hpRHa/cfZ9QbK6kVHsLssZdoCChVhXkVBCLSWkSquV73FZEHRSTar5Up26zee4zb3lpFnZphzLmnF83qRNhdklLKj7zdIvgUcIjIRcBbQCzwgd+qUrZZsfsoo2esokGtaswZ24um0dXtLkkp5WfeBoHTGFMEDAVeNsb8GWjsv7KUHX7eeYQxb6+iaXR1Zt9zCY2iwu0uSSlVDrwNgkIRGQGMBr5yjQstayYRGSQi20Vkp4hM9DC9r4hkish61+NJ70tXvvTd9sPcMXM1LevW4MOxl9CgloaAUoHC27OGbgfGAf82xuwRkVjgvbPNICLBwBSgP5AKrBaRL4wxW0s0/dEYc/051q18aOm2Q9z73lriGtbkvTsvpnaNMLtLUkqVI6+2CFxf3o8Ca13De4wxz5QxWyKw0xiz2xhTAMwGBl9Iscr3Fm4+yLj31tC+cS0+uOsSDQGlApC3Zw39AVgPLHQNdxORL8qYrSlwwG041TWupF4iskFEFohIx1Lef6yIJIlIUkZGhjclKy98tTGd+z9YS+emUcy662KiIsrc26eUqoK8PUbwFNYv/BMAxpj1WGcOnY14GGdKDK8FWhhjugKTgXmeFmSMmW6MiTfGxNevX9/LktXZfLkhnQc/XEeP5tG8e+fFRIZrCCgVqLwNgiJjTGaJcSW/1EtKBZq5DccA6WcswJgsY0y26/V8IFRE9F6HfrZm3zH+8tEG4lvU4Z07EqlZTS8wVyqQeRsEm0VkJBAsInEiMhn4pYx5VgNxIhIrImHAzcAZu5NEpJGIiOt1oqueo+f0F6hzkn4il3tmraVxdDjTb+tJRJiGgFKBztsgGA90BPKxLiTLBB4+2wyu6w4eABYB24CPjDFbRGSciIxzNfsTVshsACYBNxtjytrSUOfpVEERd7+bRH6hg7dGxxMdoQeGlVIgZX3vuk4DXWSM6Vc+JZ1dfHy8SUpKsruMSsfpNDzw4VoWbD7IjNEJXNmugd0lKaXKkYisMcbEe5pW5haBMcYBnBIRvR1VJTbp2x3M33SQx65ppyGglDqDtzuI84BNIvINkFM80hjzoF+qUj61YNOvvLxkB3/s0ZS7r2hldzlKqQrG2yD42vVQlczmtEz+56MN9GgezX+GdsZ1bF4ppU7zKgiMMe+4zvxp4xq13RhT6L+ylC9knMxn7LtJREeEMu3WnoSHBttdklKqAvIqCESkL/AOsBfrQrFmIjLaGPOD3ypTFyS/yME9s5I4dqqAT8Zdqp3IKaVK5e2uoReAAcaY7QAi0gb4EOjpr8LU+TPG8Pjczazdf4IpI3vQqake51dKlc7b6whCi0MAwBiTghfdUCt7vPnjHj5Zk8pDV8dxXRe9bYRS6uy83SJIEpG3gFmu4VHAGv+UpC7Esu2HeXrBNq7p1IiHro6zuxylVCXgbRDcC9wPPIh1jOAH4DV/FaXOz87DJ3nwg3W0axTJCzd1JShIzxBSSpXN2yAIAV4xxrwIp682rua3qtQ5O3GqgLveSaJaaBBvjI7XPoSUUl7z9hjBUsD9LubVgSW+L0edj0KHk/s/WEv6iTxev7Wn3nBeKXVOvA2C8OLuogFcryP8U5I6V//6ais/7zzKv4d2omeLOnaXo5SqZLwNghwR6VE8ICLxQK5/SlLn4v2V+3hn+T7uujyWYfHNyp5BKaVK8HZH8sPAxyKSjnVDmibAcH8VpbyzfNdR/v75Fvq0qc9j17a3uxylVCV11i0CEUkQkUbGmNVAO2AOUIR17+I95VCfKsX+o6e47/01tKgbweSR3QnWM4SUUueprF1DrwMFrte9gL8BU4DjwHQ/1qXO4mReIXe9uxqngTdHJ+j9hpVSF6SsXUPBxphjrtfDgenGmE+BT0VkvV8rUx45nIY/z1nProwc3rk9kdh6NewuSSlVyZW1RRAsIsVhcTXwrds0PVHdBv9dvJ0l2w7z5PUduDyunt3lKKWqgLK+zD8EvheRI1hnCf0IICIXYd23WJWjeevSmPrdLkYkNue2Xi3sLkcpVUWcNQiMMf8WkaVAY2Cx243lg7BuaK/KyfoDJ/jrpxtJjK3DP27oqDeYUUr5TJm7d4wxKzyMS/FPOcqTg5l5jH03iQa1qjHtlp6EhXh7+YdSSpVNv1EquLxCB2NnJZGTX8RboxOoUyPM7pKUUlWMHvCtwIwxTPhkI5vSMpl+azxtG9WyuySlVBWkWwQV2JRlO/lyQzqPDGhL/w4N7S5HKVVFaRBUUIu2HOS/i1MY3K0J9/VtbXc5SqkqTIOgAtr2axZ/nrOerjFRPHtjFz1DSCnlVxoEFczR7HzueieJWuEhTL8tnvDQYLtLUkpVcXqwuAIpKHJy73trOZKdz0f39KJhZLjdJSmlAoAGQQVhjOHJzzezau8xXrm5G12bRdtdklIqQOiuoQpi5i97mb36APdf2ZrB3ZraXY5SKoBoEFQAP6Rk8M+vttK/Q0P+0r+t3eUopQKMBoHNdmdk88AHa4lrUIuXhncjSG8wo5QqZ34NAhEZJCLbRWSniEw8S7sEEXGIyJ/8WU9Fk5lbyF3vJBESHMSbo+OpWU0P2Silyp/fgkBEgrHuZnYN0AEYISIdSmn3LLDIX7VUREUOJ+M/XMf+Y6eYOqoHzepE2F2SUipA+XOLIBHYaYzZbYwpAGYDgz20Gw98Chz2Yy0Vzn/mJ1vHBoZ04uJWde0uRykVwPwZBE2BA27Dqa5xp4lIU2AoMM2PdVQ4H60+wIyf9zDm0paMSGxudzlKqQDnzyDwdNTTlBh+GXjUGOM464JExopIkogkZWRk+Ko+W6zZd5zH523iirh6/O917e0uRyml/HpBWSrQzG04Bkgv0SYemO3qS6cecK2IFBlj5rk3MsZMB6YDxMfHlwyTSsPptC4aa1ArnFdH9CAkWE/aUkrZz59BsBqIE5FYIA24GRjp3sAYE1v8WkRmAl+VDIGq5MuN6WxJz+LFm7oSFRFqdzlKKQX4MQiMMUUi8gDW2UDBwAxjzBYRGeeaHlDHBQqKnLywOIV2jWrplcNKqQrFryeuG2PmA/NLjPMYAMaYMf6sxW6zV+9n/7FTvD0mgWC9aEwpVYHoTupykJNfxKSlO0iMrUPftvXtLkcppc6gQVAO3vxxD0eyC5h4TTu9yYxSqsLRIPCzo9n5TP9hFwM7NqRH89p2l6OUUr+jQeBnry7bSW6hgwkDtVdRpVTFpEHgRweOneK9FfsY1rMZFzWoZXc5SinlkQaBH734TQpBIjzcP87uUpRSqlQaBH6y7dcs5q1PY8xlLWkcVd3ucpRSqlQaBH7y3MJkalUL4b4+F9ldilJKnZUGgR+s2H2UZdszuLfvRdqVhFKqwtMg8DFjDM8sSKZRZDi3X9bS7nKUUqpMGgQ+tmjLIdYfOMHD/eIIDw22uxyllCqTBoEPFTmcPL8omdb1a/CnnjF2l6OUUl7RIPChT9aksisjhwkD2+q9BpRSlYZ+W/lIXqGDl5fsoFuzaAZ2bGR3OUop5TUNAh+Z+cteDmblacdySqlKR4PABzJPFfLasp30bVufS1rVtbscpZQ6JxoEPvDa9zs5mV/EXwe2s7sUpZQ6ZxoEF+jXzFxm/ryXId2a0qFJpN3lKKXUOdMguECvLNmB0xj+p38bu0tRSqnzokFwAXYePslHSQcYdXELmtWJsLscpZQ6LxoEF+D5RduJCAth/FXasZxSqvLSIDhPa/cfZ9GWQ9x9RSvq1qxmdzlKKXXeNAjOQ3HHcvVqhnHXFbF2l6OUUhdEg+A8fLc9g1V7jjH+qjhqVAuxuxyllLogGgTnyOk0PLswmeZ1IhiR2NzucpRS6oJpEJyjzzekkXzwJH8Z0IawEF19SqnKT7/JzkF+kYMXFqfQsUkkf+jSxO5ylFLKJzQIzsH7K/aTejyXRwe1IyhIO5ZTSlUNGgReOplXyKvLdnJp67pcEVfP7nKUUspnNAi89MaPeziWU8Cjg7SbaaVU1aJB4IWMk/m8+eNuru3ciK7Nou0uRymlfEqDwAuTv91BfpGTRwa0tbsUpZTyOQ2CMuw7msMHK/czPKEZrerXtLscpZTyOb8GgYgMEpHtIrJTRCZ6mD5YRDaKyHoRSRKRy/1Zz/l4YXEKIcHCQ1fH2V2KUkr5hd+CQESCgSnANUAHYISIdCjRbCnQ1RjTDbgDeNNf9ZyPzWmZfLEhnTsvj6VhZLjd5SillF/4c4sgEdhpjNltjCkAZgOD3RsYY7KNMcY1WAMwVCDPLkwmOiKUe/q0truUqsMYcDrB6QBHkfUwFeqfXamA488e05oCB9yGU4GLSzYSkaHA00AD4DpPCxKRscBYgObNy6d/n192HuHHHUd4/Nr2RIaHlst7+owxkHscTh6E7INw8pDn5/xswLi+iN2ei5eBcQ2WNu1s83mYVpqgUAiP+u1RPdr1OtrDuCgIr33muOBK9u+jVAXjzyDwdLL9774RjDFzgbki0hv4J9DPQ5vpwHSA+Ph4v/98NMbwzMJkmkSFc2uvFv5+O+85iiAn4+xf7tmHIfsQOAp+P39YTajZEGo1gsbdIDwSEBDx8EwZ0zjP+TyMK8yFvBOQl2k9ck/AiQPWuNwT4Cw8+3oJjSglNDyNcxtfPRrCakGQnjOhAps/gyAVaOY2HAOkl9bYGPODiLQWkXrGmCN+rKtM8zcdZGNqJs//qQvhocH+f8PCPOvLO/uQ61f8Ic+/5k8dAeP8/fzVa0PNRlCrIdSNs56Lh2s2sr74azaEapXwrCdjoCjPCoS8zN8HxulxJ34bl5UOh7e5xmVx9i0SsQIxPAqCq0FQMEiwFQ4S7DYcDEEhIEFnjituGxTy+3FnzFc8zdP8Ht7n9LP7o+Q4t2Epbb6yhkM0CJVfg2A1ECcisUAacDMw0r2BiFwE7DLGGBHpAYQBR/1YU5kKHU7+u3g7bRrW5I89Ynz/BimLYfOncPLX377w8078vp0EQY0G1pd5rSbQpLvbl3vDM1+HVOE7pIlAaHXrEdn43Od3OiE/67fwKDVIMq2tKKfDClunA4zjzGenA0xBiWmuts6i34/zOL/bs6dQt4unsAgOs3a7BYX+9jo41EfjS0wPCvU8XlzhWdoWpog1vdStUPHN/Bjr3+t3j9LGe2hT/Nm4kGXUbgn12/j8n99vQWCMKRKRB4BFQDAwwxizRUTGuaZPA24EbhORQiAXGO528NgWHyUdYM+RHN64LZ5gX3Ysd2QnLHoMdiyGGvWtf9C6F0HLy0v8enc916hn/YdUFyYoyNoFVD3a7kp+7/QXRJGHcCkq8XCUMVz0W9iU2aaUYeM2v6PQ2iXnKLQC0uH2unh8Qc6Z40/PU6K9cdi9pquOyx6G/v/w+WL9enstY8x8YH6JcdPcXj8LPOvPGs7FqYIiXlmyg/gWtenXvoFvFpqXBT88DyumWr9qB/wbEsdCSJhvlq8qL5HfdgdVZU6n54BwFLhCx/XaUfT78c6i3wITSj8ZofhXs8dpFzq/07VVcbaHeNkm+MKWUauhX/6J9D6Lbt7+eS+HT+YzZVSPC+9YzumEjbNhyVPWLqDut8DVf4eaPgoYpSqLoCAIqla1d2FWchoELsdzCpj23S76tW9AQss6F7awtDUw/6+QlgRN4+HmDyGmp28KVUopH9MgcHntu51kFxQxYWC7819I9mFY8g9Y/551EHfINOgyXM/KUEpVaBoEQNqJXN5Zvo8be8TQtlGtc19AUQGseh2+f846J/7SB6H3BNd5+kopVbFpEAAvfZMCwJ/7n8dpWTuWwMKJcHQHxA2AgU9DvYt8XKFSSvlPwAdByqGTfLY2lTsui6VpdHXvZzy6CxY9DikLoE5rGPkRtBnov0KVUspPAj4Inlu4nRphIdx/pZe/4vOz4cf/wvIp1gUv/f4Bl9yrZ0QopSqtgA6C1XuPsWTbISYMbEvtGmWc128MbPoYvnnSuiq46wjo95TVfYNSSlViARsExhieXZBM/VrVuP2ylmdvnL4OFjwKB1ZaXT3cNAuaJZRLnUop5W8BGwRLtx0mad9x/jWkExFhpayGnCOw9P9g7btWlw83vArdRunpoEqpKiUgg8DhNDy3KJnYejUYntDMQ4NCWP0mLHsaCnOg1/3Q569WD5VKKVXFBGQQfLY2lZRD2UwZ2YPQ4BK/7ncts04HzUiG1lfBoGegflt7ClVKqXIQcEGQV+jgpW9S6BITxbWd3Q70Ht9rnQ6a/JXVM+jNH0Lba868EYtSSlVBARcE763YR3pmHs8P62p1LFeQAz+9BD9PsnqBvPpJuOR+CNWb1SulAkNABUFWXiGvLtvJFXH1uKx1Xdj0iXU6aFYadB5mXRMQ1dTuMpVSqlwFVBC8/v0uTpwq5O8JDnj7Wtj/CzTqAje+BS162V2eUkrZImCC4HBWHp/+tIFZDb/iorlfWff5/cMr0P3Wqn9jEKWUOouACYK9K+ayMOgvRGXlWXcI6zvRCgOllApwARMEiT0TKTzYCxn0L2jQ3u5ylFKqwgiYIKBua0Jv+9TuKpRSqsLRvhKUUirAaRAopVSA0yBQSqkAp0GglFIBToNAKaUCnAaBUkoFOA0CpZQKcBoESikV4MQYY3cN50REMoB9dtdRhnrAEbuL8ILW6XuVpVat07cqQ50tjDH1PU2odEFQGYhIkjEm3u46yqJ1+l5lqVXr9K3KUmdpdNeQUkoFOA0CpZQKcBoE/jHd7gK8pHX6XmWpVev0rcpSp0d6jEAppQKcbhEopVSA0yBQSqkAp0FwnkSkmYgsE5FtIrJFRB7y0KaviGSKyHrX40mbat0rIptcNSR5mC4iMklEdorIRhHpYUONbd3W03oRyRKRh0u0sW19isgMETksIpvdxtURkW9EZIfr2eO9T0VkkIhsd63fiTbU+byIJLv+beeKSHQp8571c1IOdT4lImlu/77XljKv3etzjluNe0VkfSnzltv6vGDGGH2cxwNoDPRwva4FpAAdSrTpC3xVAWrdC9Q7y/RrgQWAAJcAK22uNxg4iHUBTIVYn0BvoAew2W3cc8BE1+uJwLOl/C27gFZAGLCh5OekHOocAIS4Xj/rqU5vPiflUOdTwCNefDZsXZ8lpr8APGn3+rzQh24RnCdjzK/GmLWu1yeBbUBTe6s6b4OBd41lBRAtIo1trOdqYJcxpsJcQW6M+QE4VmL0YOAd1+t3gCEeZk0EdhpjdhtjCoDZrvnKrU5jzGJjTJFrcAUQ46/391Yp69Mbtq/PYiIiwE3Ah/56//KiQeADItIS6A6s9DC5l4hsEJEFItKxfCs7zQCLRWSNiIz1ML0pcMBtOBV7Q+1mSv/PVRHWZ7GGxphfwfphADTw0Kairds7sLb+PCnrc1IeHnDtwppRyq62irQ+rwAOGWN2lDK9IqxPr2gQXCARqQl8CjxsjMkqMXkt1u6NrsBkYF45l1fsMmNMD+Aa4H4R6V1iuniYx5bzikUkDLgB+NjD5IqyPs9FRVq3jwNFwPulNCnrc+JvU4HWQDfgV6zdLiVVmPUJjODsWwN2r0+vaRBcABEJxQqB940xn5WcbozJMsZku17PB0JFpF45l4kxJt31fBiYi7V57S4VaOY2HAOkl091v3MNsNYYc6jkhIqyPt0cKt6F5no+7KFNhVi3IjIauB4YZVw7sEvy4nPiV8aYQ8YYhzHGCbxRyvtXlPUZAvwRmFNaG7vX57nQIDhPrv2DbwHbjDEvltKmkasdIpKItb6Pll+VICI1RKRW8WusA4ebSzT7ArjNdfbQJUBm8S4PG5T6K6sirM8SvgBGu16PBj730GY1ECcisa6tnZtd85UbERkEPArcYIw5VUobbz4nflXiuNTQUt7f9vXp0g9INsakeppYEdbnObH7aHVlfQCXY22SbgTWux7XAuOAca42DwBbsM5sWAFcakOdrVzvv8FVy+Ou8e51CjAF62yMTUC8Tes0AuuLPcptXIVYn1jh9CtQiPWr9E6gLrAU2OF6ruNq2wSY7zbvtVhnle0qXv/lXOdOrP3qxZ/TaSXrLO1zUs51znJ9/jZifbk3rojr0zV+ZvHn0q2tbevzQh/axYRSSgU43TWklFIBToNAKaUCnAaBUkoFOA0CpZQKcBoESikV4DQIlG1ExIjIC27Dj4jIUz5a9kwR+ZMvllXG+wwTqwfaZSXGt3T9fePdxr0qImPKWN44EbmtjDZjROTVUqZln0P5SgEaBMpe+cAfbb46+HdEJPgcmt8J3GeMudLDtMPAQ64Ln7xijJlmjHn3HN7fZ1xXy6oApEGg7FSEda/XP5ecUPIXffEvXbHuSfC9iHwkIiki8oyIjBKRVa6+31u7LaafiPzoane9a/5gsfrnX+3q3Owet+UuE5EPsC5qKlnPCNfyN4vIs65xT2JdWDhNRJ738PdlYF1oNrrkBBFpLSILXR2S/Sgi7VzjnxKRR1yvE1w1LnfV7H5lahPX/DtE5LkSy35BRNaKyFIRqe8a101EVshv9ySo7Rr/nYj8R0S+xwqtYa6/cYOI/ODhb1JVkAaBstsUYJSIRJ3DPF2Bh4DOwK1AG2NMIvAmMN6tXUugD3Ad1pd1ONYv+ExjTAKQANwtIrGu9olYV4B2cH8zEWmC1Y//VVgdoiWIyBBjzP8BSVj990wopdZngL942MqYDow3xvQEHgFe8zDv21hXr/YCHCWmdQOGu9bBcBEp7n+nBlZfTT2A74G/u8a/CzxqjOmCFXR/d1tWtDGmjzHmBeBJYKCxOva7oZS/SVUxGgTKVsbqsfVd4MFzmG21se4HkY/VzcBi1/hNWF/+xT4yxjiN1U3wbqAdVp8vt4l1V6mVWN1ExLnarzLG7PHwfgnAd8aYDGP16/8+1g1LvPn79gCrgJHF48TqsfZS4GNXHa9j3egItzbRQC1jzC+uUR+UWPRSY0ymMSYP2Aq0cI138ltHaO8Bl7tCNtoY871r/Dsl6nfvOO1nYKaI3I11ExgVAHSfoKoIXsbqYvptt3FFuH6ouDqac9/Pnu/22uk27OTMz3TJ/lMMVr9K440xi9wniEhfIKeU+jx1fXwu/gN8AhTvagkCThhjup1lnrLe030dOCj9/7I3fcic/ruNMeNE5GKsraj1ItLNGGNnx36qHOgWgbKdMeYY8BHWbptie4GerteDgdDzWPQwEQlyHTdoBWwHFgH3itWFOCLSxtU75NmsBPqISD3XLp4RWLtdvGKMScb61X69azgL2CMiw1w1iIh0LTHPceCkWL3BgtXLpjeCgOJjKyOBn4wxmcBxEbnCNf7W0uoXkdbGmJXGmCeBI5zZ5bOqonSLQFUUL2D1LlrsDeBzEVmFdcC1tF/rZ7Md6wuvIda+9jwReRNr99Fa15ZGBp5vMXmaMeZXEXkMWIb1S32+McZTl9Nn829gndvwKGCqiPwvVsjNxuqp0t2dwBsikgN8B2R68T45QEcRWeNqP9w1fjTWcZIIrN1kt5cy//MiEof1dy71UJOqgrT3UaUqKBGpaVw34hGRiVjdMj9kc1mqCtItAqUqrutcWyIhwD5gjL3lqKpKtwiUUirA6cFipZQKcBoESikV4DQIlFIqwGkQKKVUgNMgUEqpAPf/lsd4c8dfTkwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(num_neigh, recall_scores, label = 'Recall score')\n",
    "plt.plot(num_neigh, f1_scores, label = 'f1 score')\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel('Number of Neighbors')\n",
    "plt.legend()\n",
    "print(f'max recall score: {max(recall_scores)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is an improvement to K-NN now, as # of neightbor increases, the recall score significantly increases."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Neuro Network 2.0\n",
    "\n",
    "Previously, Neuro Network completely failed to learn the data and we received an f1-score of 0.00. We will run NN again to see if balanced data will improve the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.98      0.73      0.83     29243\n",
      "           1       0.22      0.81      0.34      2737\n",
      "\n",
      "    accuracy                           0.73     31980\n",
      "   macro avg       0.60      0.77      0.59     31980\n",
      "weighted avg       0.91      0.73      0.79     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neural_network import MLPClassifier\n",
    "nn = MLPClassifier(solver='sgd', alpha=1e-5,\n",
    "                    hidden_layer_sizes=(5, 3), random_state=1, max_iter=1000)\n",
    "nn.fit(X_train, y_train)\n",
    "y_test_pred = nn.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Neuro Network preformed the best out of all the algorithems when it comes to recall score.\n",
    "\n",
    "#### Random Forest 2.0\n",
    "Let's see if we can obtain a better result this time with Random Forest."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.95      0.94     29243\n",
      "           1       0.30      0.23      0.26      2737\n",
      "\n",
      "    accuracy                           0.89     31980\n",
      "   macro avg       0.62      0.59      0.60     31980\n",
      "weighted avg       0.88      0.89      0.88     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rf = RandomForestClassifier(criterion='gini',n_estimators=25)\n",
    "\n",
    "rf.fit(X_train, y_train)\n",
    "y_test_pred = rf.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Random Forest did have an improvement from before but actually did not preform as well as we imagined even with a much more y label and class weight balanced dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hypter-parameter tuning\n",
    "\n",
    "We are interested in fine-tuning the parameters. However, due to the limit of time, we decided to preform RandomizedSearchCV on Random Forest and Logistic Regression."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomizedSearchCV(cv=5, estimator=RandomForestClassifier(n_estimators=25),\n",
       "                   n_iter=30, n_jobs=-1,\n",
       "                   param_distributions={&#x27;criterion&#x27;: [&#x27;entropy&#x27;, &#x27;gini&#x27;],\n",
       "                                        &#x27;max_depth&#x27;: [10, 120, 230, 340, 450,\n",
       "                                                      560, 670, 780, 890,\n",
       "                                                      1000],\n",
       "                                        &#x27;max_features&#x27;: [&#x27;sqrt&#x27;, &#x27;log2&#x27;],\n",
       "                                        &#x27;min_samples_leaf&#x27;: [1, 2, 4, 6],\n",
       "                                        &#x27;min_samples_split&#x27;: [2, 5, 7],\n",
       "                                        &#x27;n_estimators&#x27;: [2, 12, 23, 34, 45, 56,\n",
       "                                                         67, 78, 89, 100]},\n",
       "                   random_state=1)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomizedSearchCV</label><div class=\"sk-toggleable__content\"><pre>RandomizedSearchCV(cv=5, estimator=RandomForestClassifier(n_estimators=25),\n",
       "                   n_iter=30, n_jobs=-1,\n",
       "                   param_distributions={&#x27;criterion&#x27;: [&#x27;entropy&#x27;, &#x27;gini&#x27;],\n",
       "                                        &#x27;max_depth&#x27;: [10, 120, 230, 340, 450,\n",
       "                                                      560, 670, 780, 890,\n",
       "                                                      1000],\n",
       "                                        &#x27;max_features&#x27;: [&#x27;sqrt&#x27;, &#x27;log2&#x27;],\n",
       "                                        &#x27;min_samples_leaf&#x27;: [1, 2, 4, 6],\n",
       "                                        &#x27;min_samples_split&#x27;: [2, 5, 7],\n",
       "                                        &#x27;n_estimators&#x27;: [2, 12, 23, 34, 45, 56,\n",
       "                                                         67, 78, 89, 100]},\n",
       "                   random_state=1)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: RandomForestClassifier</label><div class=\"sk-toggleable__content\"><pre>RandomForestClassifier(n_estimators=25)</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomForestClassifier</label><div class=\"sk-toggleable__content\"><pre>RandomForestClassifier(n_estimators=25)</pre></div></div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "RandomizedSearchCV(cv=5, estimator=RandomForestClassifier(n_estimators=25),\n",
       "                   n_iter=30, n_jobs=-1,\n",
       "                   param_distributions={'criterion': ['entropy', 'gini'],\n",
       "                                        'max_depth': [10, 120, 230, 340, 450,\n",
       "                                                      560, 670, 780, 890,\n",
       "                                                      1000],\n",
       "                                        'max_features': ['sqrt', 'log2'],\n",
       "                                        'min_samples_leaf': [1, 2, 4, 6],\n",
       "                                        'min_samples_split': [2, 5, 7],\n",
       "                                        'n_estimators': [2, 12, 23, 34, 45, 56,\n",
       "                                                         67, 78, 89, 100]},\n",
       "                   random_state=1)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "\n",
    "param_grid = {\n",
    "    'n_estimators': [int(x) for x in np.linspace(start=2,stop=100,num=10)],\n",
    "    'max_features': ['sqrt','log2'],\n",
    "    'max_depth': [int(x) for x in np.linspace(10,1000,10)],\n",
    "    'min_samples_split': [2,5,7],\n",
    "    'min_samples_leaf': [1,2,4,6],\n",
    "    'criterion': ['entropy','gini']\n",
    "}\n",
    "\n",
    "rs = RandomizedSearchCV(estimator=rf, param_distributions=param_grid, n_iter=30,\n",
    "                        cv=5, n_jobs=-1, random_state=1)\n",
    "\n",
    "rs.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameter:\n",
      "{'n_estimators': 34, 'min_samples_split': 2, 'min_samples_leaf': 1, 'max_features': 'log2', 'max_depth': 120, 'criterion': 'gini'}\n"
     ]
    }
   ],
   "source": [
    "print(\"Best Parameter:\")\n",
    "print(rs.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Random Forest 3.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.95      0.94     29243\n",
      "           1       0.30      0.21      0.25      2737\n",
      "\n",
      "    accuracy                           0.89     31980\n",
      "   macro avg       0.61      0.58      0.59     31980\n",
      "weighted avg       0.87      0.89      0.88     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rf = RandomForestClassifier(max_depth=120, max_features='log2', n_estimators=34,\n",
    "                           min_samples_split=2, min_samples_leaf=1)\n",
    "\n",
    "rf.fit(X_train, y_train)\n",
    "y_test_pred = rf.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The results of Random Forest score actually decreased. We believe that this could possibly be fixed/improved if we preform grid search. However, due to it being a lower score over all, we will not continue investigate RF further at the moment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-2 {color: black;background-color: white;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomizedSearchCV(cv=5,\n",
       "                   estimator=LogisticRegression(max_iter=1000, random_state=1),\n",
       "                   n_iter=8, n_jobs=-1,\n",
       "                   param_distributions={&#x27;C&#x27;: [0.01, 0.25, 0.5, 1, 10],\n",
       "                                        &#x27;penalty&#x27;: [&#x27;l2&#x27;],\n",
       "                                        &#x27;solver&#x27;: [&#x27;lbfgs&#x27;, &#x27;saga&#x27;]},\n",
       "                   random_state=1)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomizedSearchCV</label><div class=\"sk-toggleable__content\"><pre>RandomizedSearchCV(cv=5,\n",
       "                   estimator=LogisticRegression(max_iter=1000, random_state=1),\n",
       "                   n_iter=8, n_jobs=-1,\n",
       "                   param_distributions={&#x27;C&#x27;: [0.01, 0.25, 0.5, 1, 10],\n",
       "                                        &#x27;penalty&#x27;: [&#x27;l2&#x27;],\n",
       "                                        &#x27;solver&#x27;: [&#x27;lbfgs&#x27;, &#x27;saga&#x27;]},\n",
       "                   random_state=1)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-5\" type=\"checkbox\" ><label for=\"sk-estimator-id-5\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: LogisticRegression</label><div class=\"sk-toggleable__content\"><pre>LogisticRegression(max_iter=1000, random_state=1)</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-6\" type=\"checkbox\" ><label for=\"sk-estimator-id-6\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">LogisticRegression</label><div class=\"sk-toggleable__content\"><pre>LogisticRegression(max_iter=1000, random_state=1)</pre></div></div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "RandomizedSearchCV(cv=5,\n",
       "                   estimator=LogisticRegression(max_iter=1000, random_state=1),\n",
       "                   n_iter=8, n_jobs=-1,\n",
       "                   param_distributions={'C': [0.01, 0.25, 0.5, 1, 10],\n",
       "                                        'penalty': ['l2'],\n",
       "                                        'solver': ['lbfgs', 'saga']},\n",
       "                   random_state=1)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "logreg = LogisticRegression(max_iter=1000, random_state=1)\n",
    "\n",
    "param_grid = {\n",
    "    'penalty': ['l2'],\n",
    "    'C': [0.01, 0.25, 0.5, 1, 10],\n",
    "    'solver': ['lbfgs', 'saga']\n",
    "}\n",
    "\n",
    "rs1 = RandomizedSearchCV(estimator=logreg, param_distributions=param_grid, n_iter=8,\n",
    "                         cv=5, n_jobs=-1, random_state=1)\n",
    "\n",
    "rs1.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:\n",
      "{'solver': 'lbfgs', 'penalty': 'l2', 'C': 0.5}\n"
     ]
    }
   ],
   "source": [
    "print(\"Best Parameters:\")\n",
    "print(rs1.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Logistic Regression 3.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.97      0.75      0.85     29243\n",
      "           1       0.23      0.78      0.35      2737\n",
      "\n",
      "    accuracy                           0.75     31980\n",
      "   macro avg       0.60      0.76      0.60     31980\n",
      "weighted avg       0.91      0.75      0.80     31980\n",
      "\n"
     ]
    }
   ],
   "source": [
    "logreg = LogisticRegression(C=0.5, max_iter=1000, random_state=1,\n",
    "                           penalty='l2', solver='lbfgs')\n",
    "clf = logreg.fit(X_train, y_train)\n",
    "\n",
    "y_test_pred = clf.predict(X_test)\n",
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similarily, logistic regression with randomized search cross validation also did not have a noticable improvement. Further investigation with grid-search cross validation can be done in the future with additional parameter options such as 'l1' penalty and other solvers."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Discussion\n",
    "\n",
    "### Interpreting the result\n",
    "\n",
    "* **Logistic Regression performs best, but the model we would move forward with from this point would likely be Neuro Networks.**\n",
    "\n",
    "    With our given results, Logistic Regression outperformed the other algorithms(K-Nearest Neighbors, Neuro Networks, and Random Forest) overall when comparing both the overall recall and f1-score. However, the performance of our model is to be based on its performance primarily on how well it can identify those who do have hard disease as having hard disease. As such, while F1-score does play a factor in how we determine the model we will continue to use, both the difference between the f1-score of Logistic Regression and Neuro Networks is minimal enough to overlook in order to gain better performance in the recall for Neuro Networks.\n",
    "\n",
    "\n",
    "\n",
    "* **K-Nearest Neighbors and Random Forest both underperform in comparison to Logistic Regression and Neuro Networks.**\n",
    "    \n",
    "    Despite our attempts to increase the performance of the Random Forest model, the introduction of more y label and class weight did not make the performance of the Random Forest good enough to be questioned for further use.On the other hand, K-Nearest Neighbors did perform closely to Logistic Regression and Neuro networks, but its performance plateauted at about 72%. Further increase to the k value would only lead to minimal changes or decrease the performance as the model's decision boundary will smooth out making the model less and less complex.\n",
    "  \n",
    "\n",
    "* **Neuro Networks likely worked best due to the nature of the dataset being not linearly separable.**\n",
    "    \n",
    "    Neuro Networks works better than the other algorithms on nonlinearly separable data and since its performance in recall was the best then this was most likely the nature of the data. As shown by the performance of the K-Nearest Neighbors model, the performance of the model peaked at 72% as the k value increased. However if you were to continue to increase this value, then the performance of the algorithm would only worsen or remain relatively the same. As such the K-Nearest Neighbors's decision boundary is only getting smoother which implies that the classification of the data itself cannot be linearly separable.\n",
    "    \n",
    "    \n",
    "* **Randomized Search Cross Validation did not improve the results much**\n",
    "\n",
    "    Hoping to see better results in limited time, we went with randomized search cross validation on Logistic Regression hoping to see if it can yeld an even better result. We also ran randomized search on Random Forest, wondering if preformance could be improved. The result is somewhat dissapointing: 1) For Random Forest, the recall and f1 scores actually slightly decreased. 2) For logistic regression, there isn't a noticeable improvement. In the future, it would be great to try grid search cross validation to confirm the results.\n",
    "\n",
    "### Limitations\n",
    "\n",
    "One limitation to our work includes the limited proportions of patients diagnosed with heart disease or myocardial infarction. Having limited data or a small dataset, can lead to biases in our algorithms. We only used about 10% of our dataset for our test set, so our data size is about 32,000 observations. Another limitation we had was although we decided to use the f1-score for our evalutation metric since our data was unbalanced, however, we found that the recall score actually works better for our research question. Recall puts a heavier emphasis on False Negatives, which are very detrimental to our research question, since it would be all the patients who were not diagnosed with heart disease, but actually had a heart disease. We were worried that recall would not be a good choice for our project, because our data was so unbalanced, however, through random oversampling and data standardization, we were able to balance our data so recall would not be too biased. Therefore, recall was better suited as our evaluation metric, rather than the F1 score that is just a harmonic mean between the recall and precision. Another limitation we had was that since we had to balance our data using random oversampling and data standardization, this could lead to possible overfitting and increase the variance of our results.\n",
    "\n",
    "\n",
    "Due to the limit of time and resource, we were unable to further investigate how to improve our model preformance with parameter tuning. We tried running support vector machine, hoping that it would give us a better result than logistic regression. However, training time for SVM turns out to be 5+ hours, thus we decided to terminate the process. Our work have set some ground work for better undrestanding the data. In the future, it would be great to run cross validation and grid search to select the best features and parameters."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Ethics & Privacy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our datasets comes from a public dataset. Since our dataset comes from a survey that was collected by the CDC, the methods of data collection shouldn't be a violation of privacy as long as their name and personal information is not included. Additionally, in terms of data storage, we do not hold any personal information of the participants of the survey and it is a public dataset, so fears of data leakage or data deletion should not pose a problem. Additionally, this survey was taken by over 320,000 adults across all 50 states, so there is a very big sample to help decrease any bias and increase generalizability to adults in the U.S. However, since, only about 10% of our data includes positive cases, we would only use that smaller sample for the test set. This could lead to a potential bias towards a certain age or racial group in our data. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our main goal with this project was to see how training certain models on whether some has heart disease. Through seeing which model performed the best, it would help us understand the factors that underlie heart disease better. With our results it is possible to say that the factors that lead to heart disease may not be as simple as a linear relationship between the factor and risk of heart disease. This could be seen in our best performing model, a neural network, is knwon for being able to classify nonlinear relationships well. For research the medical field, this can serve as a introduction into understanding the factors of heart disease. In the future, people can work to better understand which classifer is the best for these factors, as well as understanding which factor matters the most in determining heart disease."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Footnotes\n",
    "<a name=\"healthlinenote\"></a>1.[^](#healthline):  Thomas, J. (16 July 2020) Heart Disease: Facts, Statistics, and You. Healthline Media. https://www.healthline.com/health/heart-disease/statistics <br>\n",
    "<a name=\"admonishnote\"></a>2.[^](#admonish): Heart Disease Facts CDC. https://www.cdc.gov/heartdisease/facts.htm <br>\n",
    "<a name=\"sotanote\"></a>3.[^](#sota): Data source: https://www.kaggle.com/datasets/kamilpytlak/personal-key-indicators-of-heart-disease <br>\n",
    "<a name=\"sotanote\"></a>4.[^](#sota): Know Your Risk for Heart Disease. CDC.: https://www.cdc.gov/heartdisease/risk_factors.htm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "3700c39376f501946acbe1871c61057b16ea95e6ed2722f303847becb1996c6c"
  },
  "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": 2
}