{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SageMaker Built-In Linear Learner Algorithm - Model for Predicting Medicare Average Hospital Spending:\n",
    "In this notebook, we show how to build a Model to predict Medicare spending per beneficiary at a hospital based on state and national average costs using Amazon SageMaker. This notebook contains code instructions for all the steps for building, training and deploying a machine learning model using SageMaker built-in Linear Learner algorithm. In this example, the notebook is created on a \"ml.c4.xlarge\" instance. Use the 'conda_python3' environment for the notebook.\n",
    "\n",
    "## Learning Objectives:\n",
    "This workshop covers both the preprocessing using SageMaker algorithms and other Python libraries. The major learnings are:\n",
    "\n",
    "1. Load data into SageMaker Notebooks\n",
    "2. Perform basic preprocessing including: feature cleaning, normalization and basic feature engineering.\n",
    "3. Perform basic feature selection/subsampling.\n",
    "4. Perform exploratory data analysis.\n",
    "5. Build, train, and deploy Linear Learner regression model.\n",
    "\n",
    "## Business Problem:\n",
    "Medicare is a national health insurance program, administered by the Center for Medicare and Medicaid Services (CMS). This is a primary health insurance for Americans who are aged 65 and older. Medicare has published historical data showing hospital’s average spending for Medicare Part A and Part B claims based on different claim types and claim periods covering 1 to 3 days prior to hospital admission up to 30 days after discharge from hospital admission. These hospital spending are price standardized and non-risk adjusted, since risk adjustment is done at the episode level of the claims spanning the entire period during the episode. The hospital average costs are listed against the corresponding state level average cost and national level average cost.\n",
    "\n",
    "In this notebook, the data is used to build a machine learning model using Amazon SageMaker built-in Linear Learner algorithm, which predicts average hospital spending cost based on the average state level spending and average national level spending. The predicted cost can be used for purposes of budget and for negotiating pricing with the hospitals. From the hospital’s perspective, the predicted average hospital spending provides visibility to claim financials that can be used by the hospitals to increase their efficiency and level of care.\n",
    "\n",
    "## Public Dataset Used:\n",
    "Medicare has published dataset showing average hospital spending on Medicare Part A and Part B claims. Both the links below refer to the same data set, one is listed in the healthdata.gov site and the other is listed at the data.medicare.gov site. The data dictionary is described in the link marked as #2 below. The dataset has hospital spending data from the year 2018 and has 67,826 data rows spanning across 13 columns. For the purposes of our analysis and machine learning, we use the dataset in csv (Comma Separated Values) format.\n",
    "1.\thttps://healthdata.gov/dataset/medicare-hospital-spending-claim\n",
    "2.\thttps://data.medicare.gov/Hospital-Compare/Medicare-Hospital-Spending-by-Claim/nrth-mfg3\n",
    "\n",
    "A direct link to download the dataset to local computer can be accessed at this link - https://data.medicare.gov/api/views/nrth-mfg3/rows.csv?accessType=DOWNLOAD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import boto3\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from pandas.plotting import scatter_matrix\n",
    "import sagemaker\n",
    "from sagemaker import get_execution_role\n",
    "from sagemaker import image_uris\n",
    "from sagemaker.deserializers import JSONDeserializer\n",
    "from sagemaker.s3 import S3Uploader, S3Downloader\n",
    "from sagemaker.serializers import CSVSerializer\n",
    "import seaborn as sn\n",
    "from sklearn.feature_selection import chi2\n",
    "from sklearn.feature_selection import SelectKBest\n",
    "from sklearn.model_selection import StratifiedShuffleSplit\n",
    "from sklearn.preprocessing import scale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sagemaker_session = sagemaker.Session()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "S3Downloader.download(\n",
    "    s3_uri=\"s3://aws-hcls-ml/workshop/immersion_day_workshop_data_DO_NOT_DELETE/data/medicare_data_07_13_2021/Medicare_Hospital_Spending_by_Claim.csv\",\n",
    "    local_path=\"data\",\n",
    "    sagemaker_session=sagemaker_session,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Preprocessing on the Raw Dataset:\n",
    "In this section we read the raw csv data set into a pandas data frame. We inspect the data using pandas head() function. We do data pre-processing using feature encoding, feature engineering, column renaming, dropping some columns that have no relevance to the prediction of `Avg_Hosp` cost and examining there are no missing values in the data set.\n",
    "\n",
    "**Note**: Many of these transformations and preprocessing steps are for demonstration purposes only and may not correspond to the optimal transformations for a specific column."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "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>Facility ID</th>\n",
       "      <th>Facility Name</th>\n",
       "      <th>State</th>\n",
       "      <th>Period</th>\n",
       "      <th>Claim Type</th>\n",
       "      <th>Avg Spending Per Episode Hospital</th>\n",
       "      <th>Avg Spending Per Episode State</th>\n",
       "      <th>Avg Spending Per Episode Nation</th>\n",
       "      <th>Percent of Spending Hospital</th>\n",
       "      <th>Percent of Spending State</th>\n",
       "      <th>Percent of Spending Nation</th>\n",
       "      <th>Start Date</th>\n",
       "      <th>End Date</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>670102</td>\n",
       "      <td>AD HOSPITAL EAST LLC</td>\n",
       "      <td>TX</td>\n",
       "      <td>1 to 3 days Prior to Index Hospital Admission</td>\n",
       "      <td>Hospice</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00%</td>\n",
       "      <td>0.00%</td>\n",
       "      <td>0.00%</td>\n",
       "      <td>01/01/2018</td>\n",
       "      <td>12/31/2018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>670102</td>\n",
       "      <td>AD HOSPITAL EAST LLC</td>\n",
       "      <td>TX</td>\n",
       "      <td>1 to 3 days Prior to Index Hospital Admission</td>\n",
       "      <td>Inpatient</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0.00%</td>\n",
       "      <td>0.03%</td>\n",
       "      <td>0.03%</td>\n",
       "      <td>01/01/2018</td>\n",
       "      <td>12/31/2018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>670102</td>\n",
       "      <td>AD HOSPITAL EAST LLC</td>\n",
       "      <td>TX</td>\n",
       "      <td>1 to 3 days Prior to Index Hospital Admission</td>\n",
       "      <td>Outpatient</td>\n",
       "      <td>48</td>\n",
       "      <td>140</td>\n",
       "      <td>152</td>\n",
       "      <td>0.17%</td>\n",
       "      <td>0.60%</td>\n",
       "      <td>0.70%</td>\n",
       "      <td>01/01/2018</td>\n",
       "      <td>12/31/2018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>670102</td>\n",
       "      <td>AD HOSPITAL EAST LLC</td>\n",
       "      <td>TX</td>\n",
       "      <td>1 to 3 days Prior to Index Hospital Admission</td>\n",
       "      <td>Skilled Nursing Facility</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00%</td>\n",
       "      <td>0.01%</td>\n",
       "      <td>0.01%</td>\n",
       "      <td>01/01/2018</td>\n",
       "      <td>12/31/2018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>670102</td>\n",
       "      <td>AD HOSPITAL EAST LLC</td>\n",
       "      <td>TX</td>\n",
       "      <td>1 to 3 days Prior to Index Hospital Admission</td>\n",
       "      <td>Durable Medical Equipment</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>0.00%</td>\n",
       "      <td>0.03%</td>\n",
       "      <td>0.04%</td>\n",
       "      <td>01/01/2018</td>\n",
       "      <td>12/31/2018</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Facility ID         Facility Name State  \\\n",
       "0       670102  AD HOSPITAL EAST LLC    TX   \n",
       "1       670102  AD HOSPITAL EAST LLC    TX   \n",
       "2       670102  AD HOSPITAL EAST LLC    TX   \n",
       "3       670102  AD HOSPITAL EAST LLC    TX   \n",
       "4       670102  AD HOSPITAL EAST LLC    TX   \n",
       "\n",
       "                                          Period                 Claim Type  \\\n",
       "0  1 to 3 days Prior to Index Hospital Admission                    Hospice   \n",
       "1  1 to 3 days Prior to Index Hospital Admission                  Inpatient   \n",
       "2  1 to 3 days Prior to Index Hospital Admission                 Outpatient   \n",
       "3  1 to 3 days Prior to Index Hospital Admission   Skilled Nursing Facility   \n",
       "4  1 to 3 days Prior to Index Hospital Admission  Durable Medical Equipment   \n",
       "\n",
       "   Avg Spending Per Episode Hospital  Avg Spending Per Episode State  \\\n",
       "0                                  0                               1   \n",
       "1                                  0                               7   \n",
       "2                                 48                             140   \n",
       "3                                  0                               3   \n",
       "4                                  0                               8   \n",
       "\n",
       "   Avg Spending Per Episode Nation Percent of Spending Hospital  \\\n",
       "0                                1                        0.00%   \n",
       "1                                7                        0.00%   \n",
       "2                              152                        0.17%   \n",
       "3                                2                        0.00%   \n",
       "4                                8                        0.00%   \n",
       "\n",
       "  Percent of Spending State Percent of Spending Nation  Start Date    End Date  \n",
       "0                     0.00%                      0.00%  01/01/2018  12/31/2018  \n",
       "1                     0.03%                      0.03%  01/01/2018  12/31/2018  \n",
       "2                     0.60%                      0.70%  01/01/2018  12/31/2018  \n",
       "3                     0.01%                      0.01%  01/01/2018  12/31/2018  \n",
       "4                     0.03%                      0.04%  01/01/2018  12/31/2018  "
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read the CSV file into panda dataframe and save it to another table so we can keep a copy of the original dataset\n",
    "# In our example we use the dataframe called table1 for all pre-processing, while the dataframe table\n",
    "# maintains a copy of the original data\n",
    "\n",
    "table = pd.read_csv(\"data/Medicare_Hospital_Spending_by_Claim.csv\")\n",
    "table1 = table.copy()\n",
    "table1.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Encode column \"State\"\n",
    "\n",
    "replace_map = {\n",
    "    \"State\": {\n",
    "        \"AK\": 1,\n",
    "        \"AL\": 2,\n",
    "        \"AR\": 3,\n",
    "        \"AZ\": 4,\n",
    "        \"CA\": 5,\n",
    "        \"CO\": 6,\n",
    "        \"CT\": 7,\n",
    "        \"DC\": 8,\n",
    "        \"DE\": 9,\n",
    "        \"FL\": 10,\n",
    "        \"GA\": 11,\n",
    "        \"HI\": 12,\n",
    "        \"IA\": 13,\n",
    "        \"ID\": 14,\n",
    "        \"IL\": 15,\n",
    "        \"IN\": 16,\n",
    "        \"KS\": 17,\n",
    "        \"KY\": 18,\n",
    "        \"LA\": 19,\n",
    "        \"MA\": 20,\n",
    "        \"ME\": 21,\n",
    "        \"MI\": 22,\n",
    "        \"MN\": 23,\n",
    "        \"MO\": 24,\n",
    "        \"MS\": 25,\n",
    "        \"MT\": 26,\n",
    "        \"NC\": 27,\n",
    "        \"ND\": 28,\n",
    "        \"NE\": 29,\n",
    "        \"NH\": 30,\n",
    "        \"NJ\": 31,\n",
    "        \"NM\": 32,\n",
    "        \"NV\": 33,\n",
    "        \"NY\": 34,\n",
    "        \"OH\": 35,\n",
    "        \"OK\": 36,\n",
    "        \"OR\": 37,\n",
    "        \"PA\": 38,\n",
    "        \"RI\": 39,\n",
    "        \"SC\": 40,\n",
    "        \"SD\": 41,\n",
    "        \"TN\": 42,\n",
    "        \"TX\": 43,\n",
    "        \"UT\": 44,\n",
    "        \"VA\": 45,\n",
    "        \"VT\": 46,\n",
    "        \"WA\": 47,\n",
    "        \"WI\": 48,\n",
    "        \"WV\": 49,\n",
    "        \"WY\": 50,\n",
    "    }\n",
    "}\n",
    "table1.replace(replace_map, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Encode column \"Period\"\n",
    "\n",
    "replace_map = {\n",
    "    \"Period\": {\n",
    "        \"1 to 3 days Prior to Index Hospital Admission\": 1,\n",
    "        \"During Index Hospital Admission\": 2,\n",
    "        \"1 through 30 days After Discharge from Index Hospital Admission\": 3,\n",
    "        \"Complete Episode\": 4,\n",
    "    }\n",
    "}\n",
    "table1.replace(replace_map, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Encode column \"Claim Type\"\n",
    "\n",
    "replace_map = {\n",
    "    \"Claim Type\": {\n",
    "        \"Home Health Agency\": 1,\n",
    "        \"Hospice\": 2,\n",
    "        \"Inpatient\": 3,\n",
    "        \"Outpatient\": 4,\n",
    "        \"Skilled Nursing Facility\": 5,\n",
    "        \"Durable Medical Equipment\": 6,\n",
    "        \"Carrier\": 7,\n",
    "        \"Total\": 8,\n",
    "    }\n",
    "}\n",
    "table1.replace(replace_map, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert the column \"Percent of Spending Hospital\tPercent of Spending\" to float, remove the percent sign and\n",
    "# divide by 100 to normalize for percentage\n",
    "\n",
    "table1[\"Percent of Spending Hospital\"] = (\n",
    "    table1[\"Percent of Spending Hospital\"].str.rstrip(\"%\").astype(\"float\")\n",
    ")\n",
    "table1[\"Percent of Spending Hospital\"] = table1[\"Percent of Spending Hospital\"] / 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert the column \"Percent of Spending State\" to float, remove the percent sign and\n",
    "# divide by 100 to normalize for percentage\n",
    "\n",
    "table1[\"Percent of Spending State\"] = (\n",
    "    table1[\"Percent of Spending State\"].str.rstrip(\"%\").astype(\"float\")\n",
    ")\n",
    "table1[\"Percent of Spending State\"] = table1[\"Percent of Spending State\"] / 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert the column \"Percent of Spending Nation\" to float, remove the percent sign and\n",
    "# divide by 100 to normalize for percentage\n",
    "\n",
    "table1[\"Percent of Spending Nation\"] = (\n",
    "    table1[\"Percent of Spending Nation\"].str.rstrip(\"%\").astype(\"float\")\n",
    ")\n",
    "table1[\"Percent of Spending Nation\"] = table1[\"Percent of Spending Nation\"] / 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop Column \"Facility Name\", Facility Id related to the facility, hence facility name is not\n",
    "# relevant for the model\n",
    "\n",
    "table1.drop([\"Facility Name\"], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Move the \"Avg Spending Per Episode Hospital\" column to the beginning, since the\n",
    "# algorithm requires the prediction column at the beginning\n",
    "\n",
    "col_name = \"Avg Spending Per Episode Hospital\"\n",
    "first_col = table1.pop(col_name)\n",
    "table1.insert(0, col_name, first_col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert integer values to float in the columns \"Avg Spending Per Episode Hospital\",\n",
    "# \"Avg Spending Per Episode State\" and \"Avg Spending Per Episode Nation\"\n",
    "# Columns with integer values are interpreted as categorical values. Changing to float avoids any mis-interpretetaion\n",
    "\n",
    "table1[\"Avg Spending Per Episode Hospital\"] = table1[\n",
    "    \"Avg Spending Per Episode Hospital\"\n",
    "].astype(\"float\")\n",
    "table1[\"Avg Spending Per Episode State\"] = table1[\n",
    "    \"Avg Spending Per Episode State\"\n",
    "].astype(\"float\")\n",
    "table1[\"Avg Spending Per Episode Nation\"] = table1[\n",
    "    \"Avg Spending Per Episode Nation\"\n",
    "].astype(\"float\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rename long column names for costs and percentage costs on the hospital, state and nation,\n",
    "# so they are easily referenced in the rest of this discussion\n",
    "\n",
    "table1.rename(\n",
    "    columns={\n",
    "        \"Avg Spending Per Episode Hospital\": \"Avg_Hosp\",\n",
    "        \"Avg Spending Per Episode State\": \"Avg_State\",\n",
    "        \"Avg Spending Per Episode Nation\": \"Avg_Nation\",\n",
    "        \"Percent of Spending Hospital\": \"Percent_Hosp\",\n",
    "        \"Percent of Spending State\": \"Percent_State\",\n",
    "        \"Percent of Spending Nation\": \"Percent_Nation\",\n",
    "    },\n",
    "    inplace=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert Start Date and End Date to datetime objects, then convert them to integers. First the data is converted\n",
    "# to Pandas datetime object. Then the year, month and days are extracted from the datetime object and\n",
    "# multipled with some weights to convert into final integer values.\n",
    "\n",
    "table1[\"Start Date\"] = pd.to_datetime(table1[\"Start Date\"])\n",
    "table1[\"End Date\"] = pd.to_datetime(table1[\"End Date\"])\n",
    "table1[\"Start Date\"] = (\n",
    "    1000 * table1[\"Start Date\"].dt.year\n",
    "    + 100 * table1[\"Start Date\"].dt.month\n",
    "    + table1[\"Start Date\"].dt.day\n",
    ")\n",
    "table1[\"End Date\"] = (\n",
    "    1000 * table1[\"End Date\"].dt.year\n",
    "    + 100 * table1[\"End Date\"].dt.month\n",
    "    + table1[\"End Date\"].dt.day\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "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>Avg_Hosp</th>\n",
       "      <th>Facility ID</th>\n",
       "      <th>State</th>\n",
       "      <th>Period</th>\n",
       "      <th>Claim Type</th>\n",
       "      <th>Avg_State</th>\n",
       "      <th>Avg_Nation</th>\n",
       "      <th>Percent_Hosp</th>\n",
       "      <th>Percent_State</th>\n",
       "      <th>Percent_Nation</th>\n",
       "      <th>Start Date</th>\n",
       "      <th>End Date</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>670102</td>\n",
       "      <td>43</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>2018101</td>\n",
       "      <td>2019231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>670102</td>\n",
       "      <td>43</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>7.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0003</td>\n",
       "      <td>0.0003</td>\n",
       "      <td>2018101</td>\n",
       "      <td>2019231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>48.0</td>\n",
       "      <td>670102</td>\n",
       "      <td>43</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>140.0</td>\n",
       "      <td>152.0</td>\n",
       "      <td>0.0017</td>\n",
       "      <td>0.0060</td>\n",
       "      <td>0.0070</td>\n",
       "      <td>2018101</td>\n",
       "      <td>2019231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>670102</td>\n",
       "      <td>43</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>2018101</td>\n",
       "      <td>2019231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>670102</td>\n",
       "      <td>43</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0003</td>\n",
       "      <td>0.0004</td>\n",
       "      <td>2018101</td>\n",
       "      <td>2019231</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Avg_Hosp  Facility ID  State  Period  Claim Type  Avg_State  Avg_Nation  \\\n",
       "0       0.0       670102     43       1           2        1.0         1.0   \n",
       "1       0.0       670102     43       1           3        7.0         7.0   \n",
       "2      48.0       670102     43       1           4      140.0       152.0   \n",
       "3       0.0       670102     43       1           5        3.0         2.0   \n",
       "4       0.0       670102     43       1           6        8.0         8.0   \n",
       "\n",
       "   Percent_Hosp  Percent_State  Percent_Nation  Start Date  End Date  \n",
       "0        0.0000         0.0000          0.0000     2018101   2019231  \n",
       "1        0.0000         0.0003          0.0003     2018101   2019231  \n",
       "2        0.0017         0.0060          0.0070     2018101   2019231  \n",
       "3        0.0000         0.0001          0.0001     2018101   2019231  \n",
       "4        0.0000         0.0003          0.0004     2018101   2019231  "
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# See the first 5 rows in the dataframe to see how the changed data looks\n",
    "\n",
    "table1.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop Columns \"Start Date\" and \"End Date\". The dataset is only for 2018, hence all start and end dates\n",
    "# are same in each row and does not impact the model\n",
    "\n",
    "table1.drop([\"Start Date\"], axis=1, inplace=True)\n",
    "table1.drop([\"End Date\"], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Avg_Hosp          0\n",
       "Facility ID       0\n",
       "State             0\n",
       "Period            0\n",
       "Claim Type        0\n",
       "Avg_State         0\n",
       "Avg_Nation        0\n",
       "Percent_Hosp      0\n",
       "Percent_State     0\n",
       "Percent_Nation    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Make sure the table do not have missing values. The following code line shows there are no missing values\n",
    "# in the table\n",
    "\n",
    "table1.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exploratory Data Analysis (EDA):\n",
    "In this section, we perform **Exploratory Data Analysis** of the data set and use various techniques for feature selection\n",
    "\n",
    "First, we see the scatter_matrix plot of the feature variables in the data frame as they relate to the prediction variable `Avg_Hosp` cost. For this we use the scatter_matrix function from pandas.plotting library.\n",
    "\n",
    "The entire dataset has 67826 data rows. For analysis, we take a random sample of 400 data rows for the scatter_matrix. Before selecting the 400 random data rows, we use the scale function from sklearn.preprocessing library to appropriately scale the values of the data columns. This helps the scatter_matrix plot label decorations fit properly. This plots helps in determining if we should keep all the feature columns while training the model.\n",
    "\n",
    "Next we use the SelectKBest class and chi2 statistical test available from sklearn.feature_selection library to find the scores of feature columns as they relate to the prediction column. This is another mechanism to determine which feature columns are relevant to keep in the model\n",
    "\n",
    "Following that, we create and visualize the correlation matrix. This is another mechanism towards feature selection prior to training the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "# After selecting the random sample of 400 data rows for the scatter_matrix analysis, this step\n",
    "# uses the scale function from sklearn.preprocessing library to scale the values. A new pandas data frame is created\n",
    "# that holds the sampled 400 data rows. We want to keep the original data set intact so we can use the original\n",
    "# data set for the subsequent training of the model\n",
    "\n",
    "table1_sample = table1.sample(n=400, random_state=2)\n",
    "standardised_table1_sample = scale(table1_sample)\n",
    "\n",
    "standardised_table1_sample = pd.DataFrame(\n",
    "    standardised_table1_sample, index=table1_sample.index, columns=table1_sample.columns\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The scatter_matrix is plotted using a tight layout for ease of visualization within the notebook. Hence,\n",
    "# in the scaled randomized sample of 400 data rows, the column names are renamed to shorter column names.\n",
    "\n",
    "standardised_table1_sample.rename(columns={\"Avg_Hosp\": \"A_Ho\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Facility ID\": \"F_Id\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"State\": \"ST\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Period\": \"Per\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Claim Type\": \"Clm\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Avg_State\": \"A_ST\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Avg_Nation\": \"A_Na\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Percent_Hosp\": \"P_Ho\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Percent_State\": \"P_ST\"}, inplace=True)\n",
    "standardised_table1_sample.rename(columns={\"Percent_Nation\": \"P_Na\"}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 792x576 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 36 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# In this step, the scatter matrix is plotted between the prediction column \"Avg_Hosp\" whose shortened name is \"A_Ho\"\n",
    "# and each of the other remaining feature columns. For clarity of visualization, we create two scatter_matrix\n",
    "# plots. The first one showing Avg_Hosp with the columns \"Facility Id\", \"State\", \"Period\" and \"Claim Type\".\n",
    "# The second plot shows Avg_Hosp relation with the columns \"Avg_State\", \"Avg_Nation\", \"Percent_Hosp\",\n",
    "# \"Percent_State\" and \"Percent_Nation\"\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "plt.figure(figsize=(11, 8))\n",
    "pd.plotting.scatter_matrix(\n",
    "    standardised_table1_sample.loc[:, \"A_Ho\":\"Clm\"], diagonal=\"kde\"\n",
    ")\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "pd.plotting.scatter_matrix(\n",
    "    standardised_table1_sample.loc[:, [\"A_Ho\", \"A_ST\", \"A_Na\", \"P_Ho\", \"P_ST\", \"P_Na\"]],\n",
    "    diagonal=\"kde\",\n",
    ")\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         Features         Score\n",
      "0     Facility ID  9.760495e+08\n",
      "4       Avg_State  7.996024e+08\n",
      "5      Avg_Nation  7.950498e+08\n",
      "1           State  8.320571e+04\n",
      "6    Percent_Hosp  3.732397e+04\n",
      "7   Percent_State  3.685061e+04\n",
      "8  Percent_Nation  3.673161e+04\n",
      "3      Claim Type  2.672140e+04\n",
      "2          Period  1.624859e+04\n"
     ]
    }
   ],
   "source": [
    "# In this step, we calculate the statistical scores of the feature columns as it relates to the prediction column\n",
    "# Avg_Hosp using the SelectKBest library function based on the chi2 statistical test. The scores are displayed in a\n",
    "# tabluar format for visualization. The X data frame has all the feature columns. The y data frame has the\n",
    "# prediction column.\n",
    "\n",
    "X = table1.iloc[:, 1:10]\n",
    "y = table1.iloc[:, 0:1]\n",
    "\n",
    "# We are selecting all the feature columns to see the scores for each feature column\n",
    "selected = SelectKBest(score_func=chi2, k=9)\n",
    "fit = selected.fit(X, y)\n",
    "datascores = pd.DataFrame(fit.scores_)\n",
    "datacolumns = pd.DataFrame(X.columns)\n",
    "\n",
    "# concat two dataframes for better visualization\n",
    "featureScores = pd.concat([datacolumns, datascores], axis=1)\n",
    "featureScores.columns = [\"Features\", \"Score\"]  # naming the dataframe columns\n",
    "print(featureScores.nlargest(9, \"Score\"))  # print 9 best features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                Avg_Hosp   Facility ID         State        Period  \\\n",
      "Avg_Hosp        1.000000 -8.863661e-06 -2.051140e-04  4.912669e-01   \n",
      "Facility ID    -0.000009  1.000000e+00  9.821420e-01 -7.250570e-16   \n",
      "State          -0.000205  9.821420e-01  1.000000e+00 -7.941509e-16   \n",
      "Period          0.491267 -7.250570e-16 -7.941509e-16  1.000000e+00   \n",
      "Claim Type      0.329614  1.675355e-14  1.650797e-14  1.815683e-01   \n",
      "Avg_State       0.978819  8.851031e-04  6.733522e-04  4.994380e-01   \n",
      "Avg_Nation      0.976734 -1.315365e-15 -1.331209e-15  5.009157e-01   \n",
      "Percent_Hosp    0.982630  1.653416e-07  1.680870e-07  5.003226e-01   \n",
      "Percent_State   0.977278  1.628740e-06  1.705365e-06  4.995657e-01   \n",
      "Percent_Nation  0.976734  1.502742e-15  1.561542e-15  5.009206e-01   \n",
      "\n",
      "                  Claim Type  Avg_State    Avg_Nation  Percent_Hosp  \\\n",
      "Avg_Hosp        3.296143e-01   0.978819  9.767342e-01  9.826304e-01   \n",
      "Facility ID     1.675355e-14   0.000885 -1.315365e-15  1.653416e-07   \n",
      "State           1.650797e-14   0.000673 -1.331209e-15  1.680870e-07   \n",
      "Period          1.815683e-01   0.499438  5.009157e-01  5.003226e-01   \n",
      "Claim Type      1.000000e+00   0.337399  3.388598e-01  3.351493e-01   \n",
      "Avg_State       3.373986e-01   1.000000  9.984595e-01  9.939270e-01   \n",
      "Avg_Nation      3.388598e-01   0.998460  1.000000e+00  9.939876e-01   \n",
      "Percent_Hosp    3.351493e-01   0.993927  9.939876e-01  1.000000e+00   \n",
      "Percent_State   3.375192e-01   0.999169  9.991468e-01  9.946930e-01   \n",
      "Percent_Nation  3.388660e-01   0.998460  1.000000e+00  9.939871e-01   \n",
      "\n",
      "                Percent_State  Percent_Nation  \n",
      "Avg_Hosp             0.977278    9.767338e-01  \n",
      "Facility ID          0.000002    1.502742e-15  \n",
      "State                0.000002    1.561542e-15  \n",
      "Period               0.499566    5.009206e-01  \n",
      "Claim Type           0.337519    3.388660e-01  \n",
      "Avg_State            0.999169    9.984596e-01  \n",
      "Avg_Nation           0.999147    1.000000e+00  \n",
      "Percent_Hosp         0.994693    9.939871e-01  \n",
      "Percent_State        1.000000    9.991469e-01  \n",
      "Percent_Nation       0.999147    1.000000e+00  \n"
     ]
    }
   ],
   "source": [
    "# Calculate Correlation Matrix to see how the data is related\n",
    "\n",
    "corrMatrix = table1.corr()\n",
    "print(corrMatrix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize Correlation matrix with Searborn and Matplotlib\n",
    "plt.subplots(figsize=(15, 10))\n",
    "plt.tick_params(labelsize=14)\n",
    "sn.heatmap(corrMatrix, annot=True, annot_kws={\"size\": 12}, fmt=\".2f\", robust=True)\n",
    "plt.xlabel(\"Parameters\", fontsize=25)\n",
    "plt.ylabel(\"Parameters\", fontsize=25)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Avg_Hosp          1.000000\n",
       "Percent_Hosp      0.982630\n",
       "Avg_State         0.978819\n",
       "Percent_State     0.977278\n",
       "Avg_Nation        0.976734\n",
       "Percent_Nation    0.976734\n",
       "Period            0.491267\n",
       "Claim Type        0.329614\n",
       "Facility ID      -0.000009\n",
       "State            -0.000205\n",
       "Name: Avg_Hosp, dtype: float64"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Our target prediction is the column Avg_Hosp, we want to see\n",
    "# how this value is correlated with the other feature columns\n",
    "# From the below matrix, we see that the prediction column Avg_Hosp has the highest correlation\n",
    "# with the Avg_State feature column\n",
    "\n",
    "corrMatrix[\"Avg_Hosp\"].sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Creating Train, Validation and Test Datasets:\n",
    "In the last step of the previous section, we see that the prediction column `Avg_Hosp` cost has second highest correlation with the `Avg_State` feature column. In this section we examine the spread of values for the `Avg_State`. Then we use a technique called stratification to categorize each data row based on the category of the `Avg_State` cost. We did not use the column `Percent_Hosp` (highest correlation) for stratification because the `Percent_Hosp` column is derived from the `Avg_Hosp` value and the line which contains the Total Cost per Episode for a specific hospital. There is an obvious expected high degree of correlation between `Avg_Hosp` and `Percent_Hosp`. Using `Percent_Hosp` to stratify values will not give us a good representative sample for train, validation and test datasets.\n",
    "\n",
    "From the defined categories of the `Avg_State` values, we use the StratifiedShuffleSplit function from the Scikit-Learn library to split and randomly select data subsets for the training, validation and test data sets. We use this method two times. First we split the original data into train and test. Then we split the test set again into validation and test set. This stratification technique allows us to have good representation of data in each of the train, validation and test sets that are well spread across the `Avg_State` value categories. Then we move the prediction column `Avg_Hosp` to the first column in each of the train, validation and test data sets. The data sets are uploaded into S3 bucket location during the machine learning model creation process."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ec2-user/anaconda3/envs/python3/lib/python3.8/site-packages/seaborn/distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZcAAAEHCAYAAABiAAtOAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAAlIklEQVR4nO3de3hd1X3m8e+ruyVLlmXLF2xjGxAYAwkhioGBXCYpBNOmTpqkgTQDTWkdJtBOp9NOnbaZSZo2QzPTMCVJIaQlD6RJCOTqNDQOIYQ0IVxsrja28Q1sY9mWLV9kyZYs+dc/zpYjZF2O5b11fMz7eR4955y119pnLR8/erX3XnsdRQRmZmZpKil0B8zM7NTjcDEzs9Q5XMzMLHUOFzMzS53DxczMUldW6A4U0uTJk2POnDmF7oaZWVFZsWLFrohoHK7Oazpc5syZw/LlywvdDTOzoiLp5ZHq+LSYmZmlzuFiZmapc7iYmVnqHC5mZpY6h4uZmaXO4WJmZqlzuJiZWeocLmZmljqHi5mZpe41fYf+yeJrj28+puyDF59egJ6YmaXDRy5mZpY6h4uZmaXO4WJmZqlzuJiZWeocLmZmljqHi5mZpS7TcJF0laS1ktZLWjLIdkm6Ldn+nKSLRmor6f2SVkk6Iqm5X/kVklZIej55fHuWYzMzs6FlFi6SSoEvAAuB+cC1kuYPqLYQaEp+FgO359F2JfBbwM8G7GsX8K6IuAC4HvhK2mMyM7P8ZHkT5QJgfURsBJB0L7AIeKFfnUXAPRERwGOS6iVNB+YM1TYiVidlr3qziHi638tVQJWkyojoymJwZmY2tCxPi80AtvR7vTUpy6dOPm2H817gaQeLmVlhZHnkokHKIs86+bQd/E2l84C/A64cYvticqfgOP10L7FiZpaFLI9ctgKz+r2eCWzLs04+bY8haSbwHeC6iNgwWJ2IuDMimiOiubGxccRBmJnZ8csyXJ4EmiTNlVQBXAMsHVBnKXBdMmvsEmBfRLTk2fZVJNUDPwA+FhG/SHksZmZ2HDILl4joAW4GlgGrgfsiYpWkGyXdmFR7ANgIrAe+BHx0uLYAkt4jaStwKfADScuSfd0MnAV8XNIzyc+UrMZnZmZDU26i1mtTc3NzLF++vNDd8JL7ZlZUJK2IiObh6vgOfTMzS53DxczMUudwMTOz1DlczMwsdQ4XMzNLncPFzMxS53AxM7PUOVzMzCx1DhczM0udw8XMzFLncDEzs9Q5XMzMLHUOFzMzS53DxczMUudwMTOz1DlczMwsdQ4XMzNLncPFzMxS53AxM7PUOVzMzCx1DhczM0udw8XMzFLncDEzs9Q5XMzMLHWZhoukqyStlbRe0pJBtkvSbcn25yRdNFJbSe+XtErSEUnNA/b3saT+WknvzHJsZmY2tMzCRVIp8AVgITAfuFbS/AHVFgJNyc9i4PY82q4Efgv42YD3mw9cA5wHXAX8Y7IfMzMbY1keuSwA1kfExojoBu4FFg2oswi4J3IeA+olTR+ubUSsjoi1g7zfIuDeiOiKiE3A+mQ/ZmY2xrIMlxnAln6vtyZl+dTJp+1o3g9JiyUtl7S8tbV1hF2amdloZBkuGqQs8qyTT9vRvB8RcWdENEdEc2Nj4wi7NDOz0SjLcN9bgVn9Xs8EtuVZpyKPtqN5PzMzGwNZHrk8CTRJmiupgtzF9qUD6iwFrktmjV0C7IuIljzbDrQUuEZSpaS55CYJPJHmgMzMLD+ZHblERI+km4FlQClwV0SsknRjsv0O4AHganIX3zuBDw/XFkDSe4DPAY3ADyQ9ExHvTPZ9H/AC0APcFBG9WY3PzMyGpoiRLmWcupqbm2P58uWF7gZfe3zzMWUfvPj0AvTEzGxkklZERPNwdXyHvpmZpc7hYmZmqXO4mJlZ6hwuZmaWOoeLmZmlzuFiZmapc7iYmVnqHC5mZpY6h4uZmaXO4WJmZqlzuJiZWeocLmZmljqHi5mZpc7hYmZmqXO4mJlZ6hwuZmaWOoeLmZmlzuFiZmapc7iYmVnqHC5mZpY6h4uZmaXO4WJmZqlzuJiZWeocLmZmlrpMw0XSVZLWSlovackg2yXptmT7c5IuGqmtpAZJD0palzxOTMrLJd0t6XlJqyV9LMuxmZnZ0DILF0mlwBeAhcB84FpJ8wdUWwg0JT+LgdvzaLsEeCgimoCHktcA7wcqI+IC4I3ARyTNyWZ0ZmY2nCyPXBYA6yNiY0R0A/cCiwbUWQTcEzmPAfWSpo/QdhFwd/L8buDdyfMAaiSVAeOAbmB/NkMzM7PhZBkuM4At/V5vTcryqTNc26kR0QKQPE5Jyr8JdAAtwGbg/0VE28BOSVosabmk5a2traMZl5mZjSDLcNEgZZFnnXzaDrQA6AVOA+YC/0PSGcfsJOLOiGiOiObGxsYRdmlmZqORZbhsBWb1ez0T2JZnneHa7khOnZE87kzKPwj8MCIOR8RO4BdAcwrjMDOz45RluDwJNEmaK6kCuAZYOqDOUuC6ZNbYJcC+5FTXcG2XAtcnz68Hvpc83wy8PdlXDXAJsCarwZmZ2dDKstpxRPRIuhlYBpQCd0XEKkk3JtvvAB4ArgbWA53Ah4drm+z6FuA+STeQC5T3J+VfAL4MrCR3Wu3LEfFcVuMzM7OhZRYuABHxALkA6V92R7/nAdyUb9ukfDfwjkHKD/CroDEzswLK67SYpG9J+nVJvqPfzMxGlG9Y3E7ugvk6SbdImpdhn8zMrMjlFS4R8eOI+B3gIuAl4EFJj0r6sKTyLDtoZmbFJ+/TXJImAb8L/D7wNPAP5MLmwUx6ZmZmRSuvC/qSvg3MA74CvKvvDnngG5KWZ9U5MzMrTvnOFvunZPbWUZIqI6IrInyjopmZvUq+p8X+ZpCyX6bZETMzO3UMe+QiaRq5BSPHSXoDv1rzqw6ozrhvZmZWpEY6LfZOchfxZwKf7VfeDvxFRn0yM7MiN2y4RMTdwN2S3hsR3xqjPpmZWZEb6bTYhyLiX4A5kv5k4PaI+OwgzczM7DVupNNiNcnj+Kw7YmZmp46RTot9MXn85Nh0x8zMTgX5Llz5GUl1ksolPSRpl6QPZd05MzMrTvne53JlROwHfoPct0SeDfxZZr0yM7Oilm+49C1OeTXw9Yhoy6g/ZmZ2Csh3+ZfvS1oDHAQ+KqkROJRdt8zMrJjlu+T+EuBSoDkiDgMdwKIsO2ZmZsXreL7m+Fxy97v0b3NPyv0xM7NTQL5L7n8FOBN4BuhNigOHi5mZDSLfI5dmYH5ERJadMTOzU0O+s8VWAtOy7IiZmZ068j1ymQy8IOkJoKuvMCJ+M5NemZlZUcs3XD6RZSfMzOzUku9U5EeAl4Dy5PmTwFMjtZN0laS1ktZLWjLIdkm6Ldn+nKSLRmorqUHSg5LWJY8T+217naRfSlol6XlJVfmMz8zM0pXv2mJ/AHwT+GJSNAP47ghtSoEvAAuB+cC1kuYPqLYQaEp+FgO359F2CfBQRDQBDyWvSaZI/wtwY0ScB7wNOJzP+MzMLF35XtC/CbgM2A8QEeuAKSO0WQCsj4iNEdEN3MuxN14uAu6JnMeAeknTR2i7CLg7eX438O7k+ZXAcxHxbNLH3RHRN23azMzGUL7h0pX8kgeOHiWMNC15BrCl3+utSVk+dYZrOzUiWgCSx76QOxsIScskPSXpfw7WKUmLJS2XtLy1tXWEIZiZ2WjkGy6PSPoLYJykK4D7ge+P0EaDlA0MpKHq5NN2oDLgcuB3ksf3SHrHMTuJuDMimiOiubGxcYRdmpnZaOQbLkuAVuB54CPAA8BfjdBmKzCr3+uZwLY86wzXdkdy6ozkcWe/fT0SEbsiojPp40WYmdmYy3e22BFyF/A/GhHvi4gv5XG3/pNAk6S5kiqAa4ClA+osBa5LZo1dAuxLTnUN13YpcH3y/Hrge8nzZcDrJFUnp+3eCryQz/jMzCxdw97nIknA/wZuJneqSpJ6gc9FxF8P1zYieiTdTO6XfilwV0SsknRjsv0OckcXVwPrgU7gw8O1TXZ9C3CfpBuAzcD7kzZ7JH2WXDAF8EBE/OC4/jXMzCwVI91E+cfkZom9KSI2AUg6A7hd0n+PiFuHaxwRD5ALkP5ld/R7HuRmouXVNinfDRxzLSXZ9i/kpiObmVkBjXRa7Drg2r5gAYiIjcCHkm1mZmbHGClcyiNi18DCiGjlV199bGZm9iojhUv3KLeZmdlr2EjXXF4vaf8g5QK8bpeZmQ1q2HCJiNKx6oiZmZ068r2J0szMLG8OFzMzS53DxczMUudwMTOz1DlcTgKHDvfy0q6OQnfDzCw1I01Ftoz940/Xc+uDL3K4N7j+0tmcM62u0F0yMzthPnIpoH0HD3Prgy8yu6GGCePKeeTFYxZDMDMrSg6XAnp4zU4O9wa/Nn8ql501mZd2d7C5rbPQ3TIzO2EOlwL64crtTK2rZObEcbxpzkTGlZfy6AYfvZhZ8XO4FMjB7l5++uJOrpw/jRKJyrJS5k2rZcPOA4z8PWxmZic3h0uB/Pu6Vg4dPsJV5087WjZncg0d3b3sOuA1Qc2suDlcCuTZrXspLRFvnD3xaNncSTUAnpZsZkXP4VIga1raObOxhqryX60NOml8BTWVZWza7XAxs+LmcCmQNdvbOXf6q+9pkcTcSdW85HAxsyLncCmAfZ2HeWXvQeYNcsPknMk17E22m5kVK4dLAazZnvv+tXOn1x6z7fSGagCe37p3LLtkZpYqh0sBrG7pC5djj1ym1FYh4IWW9jHulZlZehwuBbBmezsNNRVMqa08ZltFWQmTxleypmWwb5c2MysOmYaLpKskrZW0XtKSQbZL0m3J9uckXTRSW0kNkh6UtC55nDhgn6dLOiDpT7Mc24lYvb2dedNqkTTo9mkTqliz3UcuZla8MgsXSaXAF4CFwHzgWknzB1RbCDQlP4uB2/NouwR4KCKagIeS1/3dCvxb6gNKSUSwfkc7Z0899npLn+kTqtjc1smBrp4x7JmZWXqyPHJZAKyPiI0R0Q3cCywaUGcRcE/kPAbUS5o+QttFwN3J87uBd/ftTNK7gY3AqmyGdOLaOrrp6O5l9qTqIetMq6sCYK2PXsysSGUZLjOALf1eb03K8qkzXNupEdECkDxOAZBUA/w58MmU+p+Jl5NVj4cNlwm5cOmbVWZmVmyyDJfBLigMXJFxqDr5tB3ok8CtEXFg2E5JiyUtl7S8tbV1hF2mb0sSLn1TjgdTP66c2soy1njGmJkVqSy/iXIrMKvf65nAtjzrVAzTdoek6RHRkpxC25mUXwy8T9JngHrgiKRDEfH5/m8YEXcCdwI0NzeP+fLDm3fnwmXmxKHDRRLzptcenbJsZlZssjxyeRJokjRXUgVwDbB0QJ2lwHXJrLFLgH3Jqa7h2i4Frk+eXw98DyAi3hwRcyJiDvD/gU8PDJaTwea2TqbWVb5qTbHBzJtWx5rt7V5+38yKUmbhEhE9wM3AMmA1cF9ErJJ0o6Qbk2oPkLsAvx74EvDR4dombW4BrpC0DrgieV00Nrd1DntKrM+50+s40NXD1j1eBsbMik+Wp8WIiAfIBUj/sjv6PQ/gpnzbJuW7gXeM8L6fGEV3x8SWtk4uOXPSiPXmJUvDrNnezqw8wsjM7GTiO/THUFdPLy37D+V15HJOch+M79Q3s2LkcBlDW/ccJGL4mWJ9airLmD2p2nfqm1lRcriMoc153OPS37xpnjFmZsXJ4TKG+u5xyfcayrxpdWza3cHB7t4su2VmljqHyxjavLuTqvISGscfuxryYM6dXksErN3hU2NmVlwcLmOobxryUKshD9T3fS9rvQyMmRUZh8sYyvcelz6zJlZTXVHKai8DY2ZFxuEyRiKCLW2dx3XPSkmJOMcX9c2sCDlcxsjuZKn94zlygdypsdUt+70MjJkVFYfLGDneach9zp1Wy/5DPbTsO5RFt8zMMuFwGSP5LLU/mL6L+v5uFzMrJg6XMZLPUvuDOWdabhkYX9Q3s2LicBkj+S61P1BtVTmzGsb5or6ZFRWHyxg53mnI/c2bVscL2xwuZlY8HC5j5HinIff3uhkT2Lirg/ZDh1PulZlZNhwuY+B4ltofzAUzJwCw8hUfvZhZccj0y8Is53iW2u/ztcc3H31+oKsHgOdf2culeXzRmJlZofnIZQyM9h6XPuMry6ivLue5rfvS7JaZWWYcLmPgeJfaH8yM+nEOFzMrGg6XMXC8S+0PZubEaja3dbK3szvFnpmZZcPhMgaOd6n9wcyoHwfgoxczKwoOlzFwIve49Jk5cRwlghUv70mpV2Zm2XG4ZCwi2HwC97j0qSovZd60Op58qS2lnpmZZcfhkrHdHd10jmKp/cEsmNvA05v3crj3SAo9MzPLTqbhIukqSWslrZe0ZJDtknRbsv05SReN1FZSg6QHJa1LHicm5VdIWiHp+eTx7VmOLV8nOg25vzfNaeDg4V5WvuLrLmZ2csssXCSVAl8AFgLzgWslzR9QbSHQlPwsBm7Po+0S4KGIaAIeSl4D7ALeFREXANcDX8loaMdltEvtD+ZNcycC8MQmnxozs5NblkcuC4D1EbExIrqBe4FFA+osAu6JnMeAeknTR2i7CLg7eX438G6AiHg6IrYl5auAKkmjn/ubkpdHudT+YKbUVjF3co2vu5jZSS/LcJkBbOn3emtSlk+d4dpOjYgWgORxyiDv/V7g6YjoGrhB0mJJyyUtb21tPY7hjM6mXR2cNqHquJfaH8qCOQ08vqmNHl93MbOTWJbhMthNHQO/CH6oOvm0HfxNpfOAvwM+Mtj2iLgzIpojormxsTGfXZ6Qja0HOKNxfGr7e8vZjbQf6uHpLXtT26eZWdqyDJetwKx+r2cC2/KsM1zbHcmpM5LHnX2VJM0EvgNcFxEbUhjDCYkINrZ2cEZjTWr7vLxpMqUl4qdrd45c2cysQLIMlyeBJklzJVUA1wBLB9RZClyXzBq7BNiXnOoaru1SchfsSR6/ByCpHvgB8LGI+EWG48pb64Eu2rt6OGNyeuEyYVw5bzx9Ig+vyf6UnpnZaGUWLhHRA9wMLANWA/dFxCpJN0q6Man2ALARWA98CfjocG2TNrcAV0haB1yRvCapfxbwcUnPJD+DXY8ZMxtbOwBSPS0G8NZzGnmhZT879x9Kdb9mZmnJ9PtcIuIBcgHSv+yOfs8DuCnftkn5buAdg5T/DfA3J9jlVP0qXNI7cgH4z+dM4f8uW8tP1uzkmgWnp7pvM7M0+A79DG1sPUBlWQmnTRiX6n7PnV7L6Q3V/OD5llT3a2aWFodLhjbu6mDu5BpKSka/GvJgJPGu10/n0Q272XXgmNnWZmYF53DJUG4acrqnxPq86/Wn0Xsk+LeV2zPZv5nZiXC4ZKSrp5ctew5yxuR0L+b3OWdqLU1TxvP9ZwbO7jYzKzyHS0bW7ThA75Fg3vTaTPYviUUXnsYTL7WxaVdHJu9hZjZaDpeMrNqWW7n4vNMmZPYev908i7IS8dXHXs7sPczMRsPhkpFV2/YzvrKM2SmshjyUKXVVvPO8ady/YiuHDvdm9j5mZscr0/tcXstWbdvPudNrU50p9rXHNx9TNm1CFfsOHmbps9v47eZZg7QyMxt7PnLJQO+RYHXL/kxPifU5Y3IN86bVcscjG+g9ktfanmZmmXO4ZOCl3R10dvcy/7S6zN9LEn/49iY2tnb4pkozO2k4XDKwatt+AM4bg3ABWHj+NJqmjOfzP1nnoxczOyk4XDKw8pV9lJeKpinZTEMe6N4nt9A8p4EXdxzgT+57ZtBrM2ZmY8nhkoHHNu7mwln1VJSN3T/v+afVccbkGn60agedXT1j9r5mZoPxbLGU7es8zPOv7OOP3t40pu+bW2/sND73k3UsfW4bN7x5LlJ+M9UGO9L54MVebdnMRs9HLin75cbdRMBlZ00e8/eeWlfFO86dynNb9/H1J7aM+fubmfVxuKTs0Q27GFdeyoWz6gvy/m89u5GmKeP5xNJVPOyvQjazAnG4pOwX63exYG7DmF5v6a9E4gNvmsXZ08bzkXtW8O2ntpL7TjYzs7HjcEnRlrZONrR2cNlZkwraj+qKMr76+5fw+lkT+JP7nuW6u57gR6u2s6/z8NGgiQi6e47Q0dXj6ctmljpf0E/R/cu3IMFvvO60QneFCePKuXfxpdz96Et87ifrWPyVFQBUlJUQERzu/VWgCBhfVcaM+nGc2TieC2Zmv7KAmZ3aHC4p6T0S3L9iK29pauS0+nS/1ni0SkvE710+l/9y6Wwe27ibtdvbaW3vorRElJeWUFFWQlmJeHxTG3s6utnc1sma7e088HwLT2/ewx+8+QzeOHti3rPOzMz6OFxS8rN1rbTsO8THf2N+obtyjPLSEt7c1MiWtoPMnnTsR/5r5049+nxn+yGe3ryXxza2sWzVDi6cVc9H3nIGV543jdKUv675VDHUTauezm2vZQ6XFEQEX3xkAw01Fa/6RV2MptTmlvG//UMX8c0VW/nnn2/iv371KeZMqub333wG771oJuMqSgvdzZNK75Fg94Eudh3opqunF0nUVZXR1tFNQ01FobtnVhAOlxR8+6lXeGxjG3/7nvMLNkssbdUVZVx36Rx+5+LZLFu1nS8+soG/+u5KPvWvL3DhrHqa5zQwIzn991r8C/3IkeCpzXtY+uw2vvXUK3QMsirCP/18E5PHV3Lu9FrOm17HzIZqPnTJ7AL01mzsOVxO0Ja2Tv72gdVcdHo9177p1PslW1oirr5gOgvPn8anH1jDky+1seLlPTy+qY0ptZWcP2MCbzi9nnnTak/5azMRweqWdpY+u43vP7uNV/YepLKshKaptcybVsuU2krGlZfSeyTYd+gwLXsPsaH1AL9Yv4t/X7eL2qoy1u1o5+oLptM8p8GnGYvY4d4jvLy7g81tnWze3cmPV++ko6uHzu5eOrtzMzDrxpXTcySoKC2huqKUcRWlTKqpYOqEKqbWVjFz4jhmT6phVsM4qitOvV/FyvIeCElXAf8AlAL/FBG3DNiuZPvVQCfwuxHx1HBtJTUA3wDmAC8Bvx0Re5JtHwNuAHqBP4qIZcP1r7m5OZYvXz7q8a1u2c+Hv/wkBw/3cv+Nl3L21NEtVFlsC00e7O7lma17WfnKPl7a1UEAM+rHcdlZk7i8qZGL5zYwta6q0N1MxZEjwQst+3l4zU6WPruNdTsPUCI4a8p4Xj+znvnT66gsH/404cHuXtbuaGfVtn2s33mArp4jTB5fwX86czKXnTWJS86YxKyJ1al+sZylp62jmzXb97OmpZ3VLfv5xfpd7GjvetUU/vJSUVNZRnVFKdXlZZSVihKJEkHPkeBw7xG6e45woKuH/YeOnf7fWFvJ7IZqTp9UzeyGGmZP6nteTUNNxUn3h5ukFRHRPGydrMJFUinwInAFsBV4Erg2Il7oV+dq4A/JhcvFwD9ExMXDtZX0GaAtIm6RtASYGBF/Lmk+8HVgAXAa8GPg7IgY8vt/Rxsur+w9yN8vW8t3n3mFSeMr+coNC5g3bfTL6xdbuPTXfugw46vK+NmLrTy6YTfth3KnhxprK7lgxgTOnlrLrIZxzJpYzZS6SmqryqmtKmN8RdlJ88v0yJGgo7uHA109tHV089KuTjbtOsALLfv55Ybd7Ok8DMCCOQ1Mr6/i/NMmUFM5ur80u3p6Wbs990tqY2sH7cnptJqKUs6eVkvTlPFMq6uisa6KqbWVTKmrYmJ1OeMqSqmpKGNceelJ8+9W7CKCrp4jtB/qoaOrh70HD7N930Fa9h1ic1snL+5oZ+32A+w60HW0zeTxFdRXVzC9roppE6qYVFPBxJoKxleW5R0AEUFndy97OrvZ3dFNW0c39ePKebmtky1tnbTsO/Sq+hVlJdRUlB4Nr/nT65hYU8HE6orksTz3vLqCCdXlVJWVMK6ilKqy7P6v5BMuWR6LLQDWR8TGpDP3AouAF/rVWQTcE7mEe0xSvaTp5I5Khmq7CHhb0v5u4KfAnyfl90ZEF7BJ0vqkD79Me2AHu3tZtmo7N1w+l4+89Uwmj69M+y2KRm1VOQBvPXsKl5/VyLa9B9myp5OKshKe37qPf1/X+qp7avqrqSilrLSE0pLcX3i5v/REaYlQ8hog6Lvxk1c99jl6Yyiv3n5MuwHbITjY3UtH9+B/f8yoH8fb503l8qZJXHbmZKbUVZ3wHwKVZaW8bmY9r5tZT0Sws72LqXVVrN2+n7U72nl4bSu7DnQdM8b+qspLKC8pQcqdtsz9+/3qsaQExGszgIIgIvcZH4lIfnL/R44kZRG5Pyg6D/cOeQPxuPJSJo2vYHZDNQvmNjC1rpJpdVVH/7+fCCl3lFNTWcbMidVHy99w+kQgd8ptTxI6uzu62dPZffR0W0dXLys272Fvx+Gjf5gMp6KshKqyEirLSylNjqSU/B95x7ypfOI3zzvh8Qwly3CZAfRfPXEruaOTkerMGKHt1IhoAYiIFklT+u3rsUH29SqSFgOLk5cHJK3Nd0AD/VXyk4LJwK50dnXSKroxvgw8Cnz2+JoV3ThHwWM8Bfwc+OToxznizJQsw2WwP50G/pkwVJ182o7m/YiIO4E7R9jXmJK0fKRDzGL3WhgjvDbG6TGeOrIcZ5bzZrcCs/q9nglsy7POcG13JKfOSB77lv7N5/3MzGwMZBkuTwJNkuZKqgCuAZYOqLMUuE45lwD7klNew7VdClyfPL8e+F6/8mskVUqaCzQBT2Q1ODMzG1pmp8UiokfSzcAyctOJ74qIVZJuTLbfATxAbqbYenJTkT88XNtk17cA90m6AdgMvD9ps0rSfeQu+vcANw03U+wkc1KdpsvIa2GM8NoYp8d46shsnJne52JmZq9Np8ZaJWZmdlJxuJiZWeocLgUk6SpJayWtT1YbKCqSXpL0vKRnJC1PyhokPShpXfI4sV/9jyVjXSvpnf3K35jsZ72k21TgtS4k3SVpp6SV/cpSG1cy6eQbSfnjkuaM6QAZcoyfkPRK8nk+k6yg0betGMc4S9LDklZLWiXpvyXlp9pnOdQ4C/t5RoR/CvBDbqLCBuAMoAJ4Fphf6H4d5xheAiYPKPsMsCR5vgT4u+T5/GSMlcDcZOylybYngEvJ3av0b8DCAo/rLcBFwMosxgV8FLgjeX4N8I2TZIyfAP50kLrFOsbpwEXJ81pyS0rNPwU/y6HGWdDP00cuhXN0eZyI6Ab6lrgpdovILctD8vjufuX3RkRXRGwiN0NwgXL3KtVFxC8j9z/3nn5tCiIifga0DShOc1z99/VN4B1jfbQ2xBiHUqxjbIlkIdyIaAdWk1u141T7LIca51DGZJwOl8IZaumbYhLAjyStUG5ZHRiwPA/Qf3meoZb62TpI+ckmzXEdbRMRPcA+YFJmPT8+N0t6Ljlt1ne6qOjHmJzGeQPwOKfwZzlgnFDAz9PhUjijWeLmZHNZRFwELARukvSWYeqmudTPyWQ04zpZx3w7cCZwIdAC/H1SXtRjlDQe+BbwxxGxf7iqg5QV8zgL+nk6XAqn6JeriYhtyeNO4DvkTvUd7/I8W5PnA8tPNmmO62gbSWXABPI/RZWZiNgREb0RcQT4ErnPE4p4jJLKyf3C/WpEfDspPuU+y8HGWejP0+FSOPksj3PSklQjqbbvOXAlsJLjXJ4nOS3RLumS5Bzudf3anEzSHFf/fb0P+Elyjrug+n7hJt5D7vOEIh1j0qd/BlZHRP/FrU+pz3KocRb88xzrmQ3+edWsjavJzezYAPxloftznH0/g9yMk2eBVX39J3ce9iFgXfLY0K/NXyZjXUu/GWFAc/IffwPweZKVIwo4tq+TO41wmNxfbDekOS6gCrif3IXUJ4AzTpIxfgV4Hngu+WUyvcjHeDm5UzfPAc8kP1efgp/lUOMs6Ofp5V/MzCx1Pi1mZmapc7iYmVnqHC5mZpY6h4uZmaXO4WJmZqlzuJiZWeocLmYnQNJ7JIWkeSnu8xxJP02WSV8t6c6k/ML+y6YP0z6vemZZcriYnZhrgZ+TW2EhLbcBt0bEhRFxLvC5pPxCcjfHjSTfemaZcbiYjVKyUOBl5O5uv0bSQkn39dv+NknfT57fIOnF5IjkS5I+P8yup9NvddqIeD5ZIuivgQ8kRzQfkLRA0qOSnk4ezxmiXk2yKu6TSd1T4asd7CRXVugOmBWxdwM/jIgXJbUBu4FLJNVERAfwAeAbkk4DPk7uy7nagZ+QWzZnKLcCP5H0KPAj4MsRsVfS/wKaI+JmAEl1wFsiokfSrwGfjoj3DlLv0+TWgvo9SfXAE5J+nPTRLBM+cjEbvWvJfckbyeP7gR8C70pWjv11cgv/LQAeiYi2iDhMbo2mIUXEl4Fzk3pvAx6TVDlI1QnA/cp9VfGtwHlD7PJKYImkZ4Cfklsn6vT8hmg2Oj5yMRsFSZOAtwPnSwpyX1sdwIeBm8gtR/5kRLSP5psJI/d1BncBdyXhcf4g1T4FPBwR70m+JOqnQ3UXeG9ErD3efpiNlo9czEbnfcA9ETE7IuZExCxgE9BD7vTXHwDfSOo+AbxV0sTkiOa9w+1Y0lXJ93MgaRq5VXxfIXdKrbZf1QlJOcDv9isfWG8Z8Id9ISfpDcc5VrPj5nAxG51ryX1BWn/fIjdr7F/JfTvnvwJExCvAp8l99eyPgRfIfU3sUK4EVkp6llww/FlEbAceBub3XagHPgP8H0m/IHfk1GdgvU8B5cBzyVHQp0Y/bLP8eMl9szEgaXxEHEiOXL4D3BURA8PJ7JThIxezsfGJ5IL6SnKnz75b0N6YZcxHLmYFIukvyc0w6+/+iPjbQvTHLE0OFzMzS51Pi5mZWeocLmZmljqHi5mZpc7hYmZmqfsPDVTuxEIyT7gAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# We see the distribution of the data around the feature column \"Avg_State\" that has the highest\n",
    "# correlation to the target prediction \"Avg_Hosp\"\n",
    "# From the histogram below we see that most of the values are between 0 and 5000 for Avg_State\n",
    "\n",
    "sn.distplot(table1.Avg_State)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Before we divide this dataset into train, validation and test, we need to stratify the values of Avg_State\n",
    "# to ensure we effectively select random sets of data into the train, validation and test data sets\n",
    "# which is good representative sample based on the Avg_State values\n",
    "# To do this, we introduce a new Feature column column called \"Avg_State_Category\", we make this column\n",
    "# equal to the value of the respective Avg_State value divided by 1000 and then using the ceil (ceiling) function\n",
    "\n",
    "table1[\"Avg_State_Category\"] = np.ceil(table1[\"Avg_State\"] / 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0     14226\n",
       "1.0     35069\n",
       "2.0      6325\n",
       "3.0      3561\n",
       "4.0      2007\n",
       "5.0       472\n",
       "10.0     1060\n",
       "11.0     1749\n",
       "12.0      274\n",
       "20.0        9\n",
       "21.0      750\n",
       "22.0     1482\n",
       "23.0      544\n",
       "24.0      298\n",
       "Name: Avg_State_Category, dtype: int64"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Now see how these categories are distributed by aggregating across all the data rows\n",
    "\n",
    "table1.Avg_State_Category.value_counts().sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "# From the above distribution, we see approximately 61k out of the total 67k values have category 0.0 to 5.0\n",
    "# Only about 6k values of the total (approx 9%) of the values have category greater than 5.0\n",
    "# Hence we update the category to 6.0 for all data rows that have category greater than 5.0\n",
    "\n",
    "table1[\"Avg_State_Category\"].where(table1[\"Avg_State_Category\"] < 5, 6.0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    14226\n",
       "1.0    35069\n",
       "2.0     6325\n",
       "3.0     3561\n",
       "4.0     2007\n",
       "6.0     6638\n",
       "Name: Avg_State_Category, dtype: int64"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# We see the distribution of the values for the Avg_State_Category across the category values 0.0 to 6.0\n",
    "\n",
    "table1.Avg_State_Category.value_counts().sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We use the stratified shuffle split function available within sklearn library to create\n",
    "# train, validation and test datasets based on our defined perentages and splitting the data appropriately\n",
    "# and randomly across all the established categories with respect to the Avg_State values\n",
    "\n",
    "# The function splits the data into two parts. First we split into training set which will be 80% of the data\n",
    "# and a test set which will be 20% of the data. Then we split this test set into validation set\n",
    "# where the validation set will be 90% of the previous test set and the final test set is\n",
    "# the remaining 10% of the previous test set. This is a two fold splitting\n",
    "\n",
    "# First lets create the training set and the temporray set\n",
    "split = StratifiedShuffleSplit(n_splits=1, test_size=0.2, random_state=42)\n",
    "\n",
    "for train_index, test_index in split.split(table1, table1[\"Avg_State_Category\"]):\n",
    "    strat_train_set = table1.loc[train_index]\n",
    "    strat_test_set = table1.loc[test_index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Here we take the test set and split it into the validation set and the test set\n",
    "# as mentioned in the prior step\n",
    "split = StratifiedShuffleSplit(n_splits=1, test_size=0.1, random_state=42)\n",
    "\n",
    "for train_index, test_index in split.split(\n",
    "    strat_test_set, strat_test_set[\"Avg_State_Category\"]\n",
    "):\n",
    "    strat_validation_set = table1.loc[train_index]\n",
    "    strat_test_set = table1.loc[test_index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(67826, 54260, 12209, 1357)"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Print total rows of data from the original dataset, training set, validation set and test set\n",
    "# To see the numbers match up, add the training, validation and test data set record counts to get the total in the\n",
    "# original data set\n",
    "\n",
    "Total = table1.shape[0]\n",
    "Train = strat_train_set.shape[0]\n",
    "Validation = strat_validation_set.shape[0]\n",
    "Test = strat_test_set.shape[0]\n",
    "Total, Train, Validation, Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop the column \"Avg_State_category\" from train, validation and test datasets. This column\n",
    "# was introduced to do stratification of the data for meaningful sampling and is not relevant\n",
    "# for the model anymore\n",
    "\n",
    "strat_train_set.drop([\"Avg_State_Category\"], axis=1, inplace=True)\n",
    "strat_validation_set.drop([\"Avg_State_Category\"], axis=1, inplace=True)\n",
    "strat_test_set.drop([\"Avg_State_Category\"], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Replace the bucket name with the your bucket name obtained form the CloudFormation output tab\n",
    "# From the SageMaker library, we use the get_execution_role function to get the execution\n",
    "# role for SageMaker to access AWS reqources while creating the machine learning model\n",
    "\n",
    "# Define IAM role\n",
    "role = get_execution_role()\n",
    "bucket = sagemaker_session.default_bucket()\n",
    "prefix = \"linear_learner\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split each of the Train, Validation and Test datasets from pandas data frame into two subsets.\n",
    "# The x subset is the feature columns and the y subset is the label column\n",
    "\n",
    "x_train = strat_train_set.iloc[:, 1:10]\n",
    "y_train = strat_train_set.iloc[:, 0:1]\n",
    "\n",
    "x_validation = strat_validation_set.iloc[:, 1:10]\n",
    "y_validation = strat_validation_set.iloc[:, 0:1]\n",
    "\n",
    "x_test = strat_test_set.iloc[:, 1:10]\n",
    "y_test = strat_test_set.iloc[:, 0:1]\n",
    "\n",
    "train_df = pd.concat([y_train, x_train], axis=1)\n",
    "validation_df = pd.concat([y_validation, x_validation], axis=1)\n",
    "test_df = pd.concat([y_test, x_test], axis=1)\n",
    "\n",
    "# copy the training dataframe to s3\n",
    "train_df.to_csv(\"data/train_data.csv\", index=False, header=False)\n",
    "validation_df.to_csv(\"data/validation_data.csv\", index=False, header=False)\n",
    "test_df.to_csv(\"data/test_data.csv\", index=False, header=False)\n",
    "\n",
    "train_data_location = f\"s3://{bucket}/{prefix}/data/train\"\n",
    "validation_data_location = f\"s3://{bucket}/{prefix}/data/validation\"\n",
    "test_data_location = f\"s3://{bucket}/{prefix}/data/test\"\n",
    "\n",
    "# Upload the training data to S3\n",
    "S3Uploader.upload(\n",
    "    local_path=\"data/train_data.csv\",\n",
    "    desired_s3_uri=train_data_location,\n",
    "    sagemaker_session=sagemaker_session,\n",
    ")\n",
    "\n",
    "# Upload the validation data to S3\n",
    "S3Uploader.upload(\n",
    "    local_path=\"data/validation_data.csv\",\n",
    "    desired_s3_uri=validation_data_location,\n",
    "    sagemaker_session=sagemaker_session,\n",
    ")\n",
    "\n",
    "# Upload the testing data to S3\n",
    "S3Uploader.upload(\n",
    "    local_path=\"data/test_data.csv\",\n",
    "    desired_s3_uri=test_data_location,\n",
    "    sagemaker_session=sagemaker_session,\n",
    ")\n",
    "\n",
    "\n",
    "output_location = f\"s3://{bucket}/{prefix}/output\"\n",
    "\n",
    "train_data_location_input = sagemaker.inputs.TrainingInput(\n",
    "    train_data_location,\n",
    "    distribution=\"FullyReplicated\",\n",
    "    content_type=\"text/csv\",\n",
    "    s3_data_type=\"S3Prefix\",\n",
    "    record_wrapping=None,\n",
    "    compression=None,\n",
    ")\n",
    "\n",
    "validation_data_location_input = sagemaker.inputs.TrainingInput(\n",
    "    validation_data_location,\n",
    "    distribution=\"FullyReplicated\",\n",
    "    content_type=\"text/csv\",\n",
    "    s3_data_type=\"S3Prefix\",\n",
    "    record_wrapping=None,\n",
    "    compression=None,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Machine Learning Model using Amazon SageMaker:\n",
    "In this section, the SageMaker built-in Linear Learner algorithm is used to train the model using the training and validation data sets as input channels. The algorithm is used in the “Regressor” mode to train the model. The boto3 Python library for AWS and the SageMaker library for Python is used. In the step below, replace with your own **bucket name** from CloudFormation Outputs tab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "# In this step, we initialize the Linear Learner Estimator\n",
    "# We define the parameters for the estimator which will be used by SageMaker to train the model.\n",
    "# In this example, we use an instance of type \"ml.c4.xlarge\". Note that the SageMaker built-in Linear Learner\n",
    "# algorithm do not need GPU type instances mandatorily. GPU instances can be selected if the dataset is large\n",
    "# and using GPU instances will help boost the performance of the model creation process.\n",
    "\n",
    "# from sagemaker import LinearLearner\n",
    "# from sagemaker.sklearn.estimator import LinearLearner\n",
    "\n",
    "\n",
    "container = image_uris.retrieve(\n",
    "    region=boto3.Session().region_name, framework=\"linear-learner\"\n",
    ")\n",
    "\n",
    "sess = sagemaker.Session()\n",
    "\n",
    "linear = sagemaker.estimator.Estimator(\n",
    "    container,\n",
    "    role,\n",
    "    instance_count=1,\n",
    "    instance_type=\"ml.c4.xlarge\",\n",
    "    output_path=output_location,\n",
    "    sagemaker_session=sagemaker_session,\n",
    "    disable_profiler=True,\n",
    ")\n",
    "\n",
    "linear.set_hyperparameters(predictor_type=\"regressor\", mini_batch_size=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'DataSource': {'S3DataSource': {'S3DataType': 'S3Prefix',\n",
       "   'S3Uri': 's3://sagemaker-ap-southeast-2-431579215499/linear_learner/data/validation',\n",
       "   'S3DataDistributionType': 'FullyReplicated'}},\n",
       " 'ContentType': 'text/csv'}"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validation_data_location_input.config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'DataSource': {'S3DataSource': {'S3DataType': 'S3Prefix',\n",
       "   'S3Uri': 's3://sagemaker-ap-southeast-2-431579215499/linear_learner/data/train',\n",
       "   'S3DataDistributionType': 'FullyReplicated'}},\n",
       " 'ContentType': 'text/csv'}"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data_location_input.config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "2022-06-06 01:44:54 Starting - Starting the training job.....\n",
      "2022-06-06 01:45:20 Starting - Preparing the instances for training.............\n",
      "2022-06-06 01:46:33 Downloading - Downloading input data..\n",
      "2022-06-06 01:46:49 Training - Downloading the training image.........\n",
      "2022-06-06 01:47:39 Training - Training image download completed. Training in progress..............................\n",
      "2022-06-06 01:50:10 Uploading - Uploading generated training model.\n",
      "2022-06-06 01:50:16 Completed - Training job completed\n"
     ]
    }
   ],
   "source": [
    "# In this step, we call the fit function to train the model using the training dataset and the validation dataset\n",
    "# llearner.fit([train_data_location,validation_data_location])\n",
    "linear.fit(\n",
    "    inputs={\n",
    "        \"train\": train_data_location_input,\n",
    "        \"validation\": validation_data_location_input,\n",
    "    },\n",
    "    logs=\"None\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------!"
     ]
    }
   ],
   "source": [
    "# In this step, we deploy the model created in the previous step as an endpoint. In this example,\n",
    "# we use an instance type of \"ml.m4.xlarge\" to deploy the model. Once deployed, the endpoint\n",
    "# can be invoked to make inference and predict the value of the \"Avg_Hosp\" cost. Please note deploying the\n",
    "# model to an endpoint takes a few minutes\n",
    "\n",
    "llearner_predictor = linear.deploy(\n",
    "    initial_instance_count=1, instance_type=\"ml.m4.xlarge\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5.0107e+04, 5.0000e+00, 2.0000e+00, 4.0000e+00, 0.0000e+00,\n",
       "       0.0000e+00, 0.0000e+00, 0.0000e+00, 0.0000e+00])"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test.values[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'linear-learner-2022-06-06-00-56-35-125'"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "llearner_predictor.endpoint_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "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>Facility ID</th>\n",
       "      <th>State</th>\n",
       "      <th>Period</th>\n",
       "      <th>Claim Type</th>\n",
       "      <th>Avg_State</th>\n",
       "      <th>Avg_Nation</th>\n",
       "      <th>Percent_Hosp</th>\n",
       "      <th>Percent_State</th>\n",
       "      <th>Percent_Nation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5380</th>\n",
       "      <td>50107</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2823</th>\n",
       "      <td>30103</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0004</td>\n",
       "      <td>0.0003</td>\n",
       "      <td>0.0004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12310</th>\n",
       "      <td>90001</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9535</th>\n",
       "      <td>50701</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>585.0</td>\n",
       "      <td>585.0</td>\n",
       "      <td>0.0312</td>\n",
       "      <td>0.0262</td>\n",
       "      <td>0.0270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10433</th>\n",
       "      <td>50779</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1923</th>\n",
       "      <td>20018</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12701</th>\n",
       "      <td>100025</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>72.0</td>\n",
       "      <td>152.0</td>\n",
       "      <td>0.0067</td>\n",
       "      <td>0.0033</td>\n",
       "      <td>0.0070</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1579</th>\n",
       "      <td>10139</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>198.0</td>\n",
       "      <td>149.0</td>\n",
       "      <td>0.0079</td>\n",
       "      <td>0.0095</td>\n",
       "      <td>0.0069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5863</th>\n",
       "      <td>50138</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>10415.0</td>\n",
       "      <td>10139.0</td>\n",
       "      <td>0.5356</td>\n",
       "      <td>0.4669</td>\n",
       "      <td>0.4684</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7136</th>\n",
       "      <td>50296</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0004</td>\n",
       "      <td>0.0003</td>\n",
       "      <td>0.0004</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1357 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Facility ID  State  Period  Claim Type  Avg_State  Avg_Nation  \\\n",
       "5380         50107      5       2           4        0.0         0.0   \n",
       "2823         30103      4       1           6        8.0         8.0   \n",
       "12310        90001      8       2           2        0.0         0.0   \n",
       "9535         50701      5       1           7      585.0       585.0   \n",
       "10433        50779      5       1           2        0.0         1.0   \n",
       "...            ...    ...     ...         ...        ...         ...   \n",
       "1923         20018      1       2           1        0.0         0.0   \n",
       "12701       100025     10       1           4       72.0       152.0   \n",
       "1579         10139      2       3           2      198.0       149.0   \n",
       "5863         50138      5       2           3    10415.0     10139.0   \n",
       "7136         50296      5       1           6        8.0         8.0   \n",
       "\n",
       "       Percent_Hosp  Percent_State  Percent_Nation  \n",
       "5380         0.0000         0.0000          0.0000  \n",
       "2823         0.0004         0.0003          0.0004  \n",
       "12310        0.0000         0.0000          0.0000  \n",
       "9535         0.0312         0.0262          0.0270  \n",
       "10433        0.0000         0.0000          0.0000  \n",
       "...             ...            ...             ...  \n",
       "1923         0.0000         0.0000          0.0000  \n",
       "12701        0.0067         0.0033          0.0070  \n",
       "1579         0.0079         0.0095          0.0069  \n",
       "5863         0.5356         0.4669          0.4684  \n",
       "7136         0.0004         0.0003          0.0004  \n",
       "\n",
       "[1357 rows x 9 columns]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "# In this step, we invoke the endpoint for a single inference. The first data row from the Test dataset is passed\n",
    "# to the endpoint for prediction. The predicted value is returned in the key value pair.\n",
    "\n",
    "# result = llearner_predictor.predict(x_test.values[0].astype('float32'))\n",
    "# print(result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "llearner_predictor.serializer = CSVSerializer()\n",
    "llearner_predictor.deserializer = JSONDeserializer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'predictions': [{'score': 55.096778869628906}]}"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "llearner_predictor.predict(\n",
    "    x_test.iloc()[0].tolist(), initial_args={\"ContentType\": \"text/csv\"}\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_predictions = [\n",
    "    llearner_predictor.predict(\n",
    "        x_test.iloc()[i].tolist(), initial_args={\"ContentType\": \"text/csv\"}\n",
    "    )\n",
    "    for i in range(0, x_test.shape[0])\n",
    "]\n",
    "all_predictions_scores = [\n",
    "    all_predictions[i][\"predictions\"][0][\"score\"]\n",
    "    for i in range(0, len(all_predictions))\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fa4aa3ca370>"
      ]
     },
     "execution_count": 49,
     "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": [
    "matplotlib.pyplot.scatter(y_test[\"Avg_Hosp\"], all_predictions_scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once you are finished testing, clean up the endpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "llearner_predictor.delete_endpoint()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Summary\n",
    "We see how the SageMaker built-in Linear Learner algorithm is used to train machine learning model and use this model for inference. The steps show how Jupyter notebooks in SageMaker can be used for build, train and deployment of machine learning models and evaluation of metrics from the model’s performance. This approach can be used in a wide variety of use cases at scale."
   ]
  }
 ],
 "metadata": {
  "instance_type": "ml.m5.4xlarge",
  "kernelspec": {
   "display_name": "conda_python3",
   "language": "python",
   "name": "conda_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": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}