A3-Used-Car-Price-Predictions.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Copyright © 2021, Chloe Wang"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Summary of Results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Training accuracy**: 0.90239\\\n",
    "**Test accuracy**: 0.90073"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Question 1**\\\n",
    "Uses training data to predict whether a used car is <$12,000, and then uses available features to predict this constructed binary variable (codes below). \n",
    "\n",
    "**Question 2**\\\n",
    "Illustrates the most important features in constructing our predictions:\\\n",
    "![A3_ImportantFeats](A3_ImportantFeatures.png)\n",
    "\n",
    "\n",
    "**Question 3**\\\n",
    "A graph that shows the training and test error-flexibility tradeoff (codes below).\\\n",
    "![A3_ErrorFlexTradeoff](A3_ErrorFlexibilityTradeoff.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Confusion Matrix of Training Data\n",
    "![A3_CMTrain](A3_ConfusionMatrix_Train.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Confusion Matrix for Test Data\n",
    "![A3_CMTest](A3_ConfusionMatrix_Test.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "ZYL5dtzG-zXT"
   },
   "outputs": [],
   "source": [
    "# importing commonly used libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# libaries for modelling and preprocessing\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split, cross_val_score, cross_validate\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "id": "k76175qJ-9m3"
   },
   "outputs": [],
   "source": [
    "# reading the two csv files\n",
    "small_df = pd.read_csv('6Z. final_predcomp_training_data_small.csv')\n",
    "large_df = pd.read_csv('6Z. final_predcomp_training_data_large.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 288
    },
    "id": "d-m6OPlc_0hx",
    "outputId": "3d5b140e-ec49-486e-c484-4a731eb18707"
   },
   "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>id</th>\n",
       "      <th>price</th>\n",
       "      <th>year</th>\n",
       "      <th>odometer</th>\n",
       "      <th>cylinders</th>\n",
       "      <th>description_length</th>\n",
       "      <th>description_credit</th>\n",
       "      <th>description_owner</th>\n",
       "      <th>description_badcredit</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>2.999900e+04</td>\n",
       "      <td>29999.000000</td>\n",
       "      <td>29933.000000</td>\n",
       "      <td>26728.000000</td>\n",
       "      <td>19101.000000</td>\n",
       "      <td>29217.000000</td>\n",
       "      <td>29999.000000</td>\n",
       "      <td>29999.000000</td>\n",
       "      <td>29999.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>7.235216e+09</td>\n",
       "      <td>15879.012534</td>\n",
       "      <td>2010.724184</td>\n",
       "      <td>94720.290482</td>\n",
       "      <td>5.890425</td>\n",
       "      <td>2352.681521</td>\n",
       "      <td>0.413047</td>\n",
       "      <td>0.187906</td>\n",
       "      <td>0.155205</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>4.594456e+06</td>\n",
       "      <td>11152.826912</td>\n",
       "      <td>8.579819</td>\n",
       "      <td>61748.952567</td>\n",
       "      <td>1.609066</td>\n",
       "      <td>2120.058619</td>\n",
       "      <td>0.492389</td>\n",
       "      <td>0.390644</td>\n",
       "      <td>0.362106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>7.213844e+09</td>\n",
       "      <td>2000.000000</td>\n",
       "      <td>1923.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>7.231904e+09</td>\n",
       "      <td>6900.000000</td>\n",
       "      <td>2008.000000</td>\n",
       "      <td>42790.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>437.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>7.236398e+09</td>\n",
       "      <td>12990.000000</td>\n",
       "      <td>2013.000000</td>\n",
       "      <td>90656.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>1808.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>7.239320e+09</td>\n",
       "      <td>22990.000000</td>\n",
       "      <td>2016.000000</td>\n",
       "      <td>135740.250000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>4140.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>7.241018e+09</td>\n",
       "      <td>50000.000000</td>\n",
       "      <td>2021.000000</td>\n",
       "      <td>500000.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>9994.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 id         price          year       odometer     cylinders  \\\n",
       "count  2.999900e+04  29999.000000  29933.000000   26728.000000  19101.000000   \n",
       "mean   7.235216e+09  15879.012534   2010.724184   94720.290482      5.890425   \n",
       "std    4.594456e+06  11152.826912      8.579819   61748.952567      1.609066   \n",
       "min    7.213844e+09   2000.000000   1923.000000       0.000000      3.000000   \n",
       "25%    7.231904e+09   6900.000000   2008.000000   42790.000000      4.000000   \n",
       "50%    7.236398e+09  12990.000000   2013.000000   90656.000000      6.000000   \n",
       "75%    7.239320e+09  22990.000000   2016.000000  135740.250000      8.000000   \n",
       "max    7.241018e+09  50000.000000   2021.000000  500000.000000     12.000000   \n",
       "\n",
       "       description_length  description_credit  description_owner  \\\n",
       "count        29217.000000        29999.000000       29999.000000   \n",
       "mean          2352.681521            0.413047           0.187906   \n",
       "std           2120.058619            0.492389           0.390644   \n",
       "min              0.000000            0.000000           0.000000   \n",
       "25%            437.000000            0.000000           0.000000   \n",
       "50%           1808.000000            0.000000           0.000000   \n",
       "75%           4140.000000            1.000000           0.000000   \n",
       "max           9994.000000            1.000000           1.000000   \n",
       "\n",
       "       description_badcredit  \n",
       "count           29999.000000  \n",
       "mean                0.155205  \n",
       "std                 0.362106  \n",
       "min                 0.000000  \n",
       "25%                 0.000000  \n",
       "50%                 0.000000  \n",
       "75%                 0.000000  \n",
       "max                 1.000000  "
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# looking at the summary statistics\n",
    "small_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "gZQo9zh7AZt8",
    "outputId": "2b0f648b-6616-4e36-ba00-64bc1047f06c"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                           0\n",
       "price                        0\n",
       "year                        66\n",
       "odometer                  3271\n",
       "cylinders                10898\n",
       "transmission               126\n",
       "drive                     8486\n",
       "type                      7094\n",
       "paint_color               8585\n",
       "size                     21039\n",
       "title_status               140\n",
       "condition                11750\n",
       "fuel                       210\n",
       "state                      403\n",
       "manufacturer               973\n",
       "description_length         782\n",
       "description_credit           0\n",
       "description_owner            0\n",
       "description_badcredit        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# missing values in the small dataset\n",
    "small_df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 288
    },
    "id": "DQ03bd7PAUin",
    "outputId": "3f0be080-f61f-4bd5-936f-288c5efa48d0"
   },
   "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>id</th>\n",
       "      <th>price</th>\n",
       "      <th>year</th>\n",
       "      <th>odometer</th>\n",
       "      <th>cylinders</th>\n",
       "      <th>description_length</th>\n",
       "      <th>description_credit</th>\n",
       "      <th>description_owner</th>\n",
       "      <th>description_badcredit</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3.000000e+05</td>\n",
       "      <td>300000.000000</td>\n",
       "      <td>299376.000000</td>\n",
       "      <td>267456.000000</td>\n",
       "      <td>191090.000000</td>\n",
       "      <td>292201.000000</td>\n",
       "      <td>300000.000000</td>\n",
       "      <td>300000.000000</td>\n",
       "      <td>300000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>7.235226e+09</td>\n",
       "      <td>15863.518507</td>\n",
       "      <td>2010.629423</td>\n",
       "      <td>94386.056742</td>\n",
       "      <td>5.888874</td>\n",
       "      <td>2379.174883</td>\n",
       "      <td>0.412167</td>\n",
       "      <td>0.186263</td>\n",
       "      <td>0.155067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>4.595326e+06</td>\n",
       "      <td>11112.310328</td>\n",
       "      <td>8.771992</td>\n",
       "      <td>61497.290868</td>\n",
       "      <td>1.596020</td>\n",
       "      <td>2146.833466</td>\n",
       "      <td>0.492226</td>\n",
       "      <td>0.389320</td>\n",
       "      <td>0.361969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>7.208550e+09</td>\n",
       "      <td>2000.000000</td>\n",
       "      <td>1900.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>7.231949e+09</td>\n",
       "      <td>6900.000000</td>\n",
       "      <td>2008.000000</td>\n",
       "      <td>42971.750000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>439.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>7.236401e+09</td>\n",
       "      <td>12991.000000</td>\n",
       "      <td>2013.000000</td>\n",
       "      <td>90054.500000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>1828.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>7.239318e+09</td>\n",
       "      <td>22900.000000</td>\n",
       "      <td>2016.000000</td>\n",
       "      <td>135166.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>4160.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>7.241019e+09</td>\n",
       "      <td>50000.000000</td>\n",
       "      <td>2021.000000</td>\n",
       "      <td>500000.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>9999.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 id          price           year       odometer  \\\n",
       "count  3.000000e+05  300000.000000  299376.000000  267456.000000   \n",
       "mean   7.235226e+09   15863.518507    2010.629423   94386.056742   \n",
       "std    4.595326e+06   11112.310328       8.771992   61497.290868   \n",
       "min    7.208550e+09    2000.000000    1900.000000       0.000000   \n",
       "25%    7.231949e+09    6900.000000    2008.000000   42971.750000   \n",
       "50%    7.236401e+09   12991.000000    2013.000000   90054.500000   \n",
       "75%    7.239318e+09   22900.000000    2016.000000  135166.000000   \n",
       "max    7.241019e+09   50000.000000    2021.000000  500000.000000   \n",
       "\n",
       "           cylinders  description_length  description_credit  \\\n",
       "count  191090.000000       292201.000000       300000.000000   \n",
       "mean        5.888874         2379.174883            0.412167   \n",
       "std         1.596020         2146.833466            0.492226   \n",
       "min         3.000000            0.000000            0.000000   \n",
       "25%         4.000000          439.000000            0.000000   \n",
       "50%         6.000000         1828.000000            0.000000   \n",
       "75%         8.000000         4160.000000            1.000000   \n",
       "max        12.000000         9999.000000            1.000000   \n",
       "\n",
       "       description_owner  description_badcredit  \n",
       "count      300000.000000          300000.000000  \n",
       "mean            0.186263               0.155067  \n",
       "std             0.389320               0.361969  \n",
       "min             0.000000               0.000000  \n",
       "25%             0.000000               0.000000  \n",
       "50%             0.000000               0.000000  \n",
       "75%             0.000000               0.000000  \n",
       "max             1.000000               1.000000  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# looking at the summary statistics\n",
    "large_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "uwuf_G4LAfHR",
    "outputId": "d3b86b77-04ca-4f43-a36f-4573d898de2a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                            0\n",
       "price                         0\n",
       "year                        624\n",
       "odometer                  32544\n",
       "cylinders                108910\n",
       "transmission               1344\n",
       "drive                     85050\n",
       "type                      71880\n",
       "paint_color               85871\n",
       "size                     210274\n",
       "title_status               1398\n",
       "condition                117743\n",
       "fuel                       1931\n",
       "state                      3787\n",
       "manufacturer              10045\n",
       "description_length         7799\n",
       "description_credit            0\n",
       "description_owner             0\n",
       "description_badcredit         0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# missing values in the large dataset\n",
    "large_df.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "wGdJ41G7gUCg"
   },
   "source": [
    "# Defining Preprocessing Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "id": "dWvpzCzAii1-"
   },
   "outputs": [],
   "source": [
    "def missing_value_impute(df):\n",
    "  # let's create a dummy variable column for missing values by iterating through all columns\n",
    "  # creating list to contain missing row indices\n",
    "  missing_row = []\n",
    "\n",
    "  # iterates through each row of large dataset\n",
    "  for index, row in df.iterrows():\n",
    "      \n",
    "      # checks each column to see if there are any missing values\n",
    "      missing_series = row.isnull()\n",
    "\n",
    "      # if any of the columns return true, add to our missing row list\n",
    "      if missing_series.any():\n",
    "          missing_row.append(index)\n",
    "  \n",
    "  # adding the dummy variable column\n",
    "  df.loc[missing_row, 'missing_value'] = 1\n",
    "\n",
    "  # fill the dummy variable columns w/ 0 for rows w/ non-missing data\n",
    "  df['missing_value'] = df['missing_value'].fillna(0)\n",
    "\n",
    "  # creating a list for categorical and numerical values\n",
    "  num_col = ['price', 'year', 'odometer', 'cylinders', 'description_length']\n",
    "\n",
    "  cat_col = ['transmission', 'drive', 'type', 'paint_color',\n",
    "            'size', 'title_status', 'condition', 'fuel',\n",
    "            'state', 'manufacturer', 'description_credit',\n",
    "             'description_owner', 'description_badcredit', 'missing_value']\n",
    "\n",
    "  # imputing the missing values using the mean for numerical columns and mode for categorical\n",
    "  # use for loop to loop through all columns\n",
    "  for col in df.columns[1:]:\n",
    "    if col in num_col:\n",
    "      df[col] = df[col].fillna(df[col].mean())\n",
    "    elif col in cat_col:\n",
    "      df[col] = df[col].fillna(df[col].mode().values[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "id": "3Tt9Prisjg3T"
   },
   "outputs": [],
   "source": [
    "missing_value_impute(large_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "id": "M7X8MVKr4OtZ"
   },
   "outputs": [],
   "source": [
    "# construct dummy variable for below $12,000\n",
    "large_df['price_target'] = large_df['price'].apply(lambda x: 1 if x < 12000 else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "id": "Y892LR9lhebL"
   },
   "outputs": [],
   "source": [
    "# list of numerical and categorical columns\n",
    "num_col = ['year', 'odometer', 'cylinders', 'description_length']\n",
    "\n",
    "cat_col = ['transmission', 'drive', 'type', 'paint_color',\n",
    "          'size', 'title_status', 'condition', 'fuel',\n",
    "          'state', 'manufacturer', 'description_credit',\n",
    "            'description_owner', 'description_badcredit', 'missing_value']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "r6Rx0Y03-8rN",
    "outputId": "432dd2c2-82d3-4b69-b38d-819a87f30424"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['transmission',\n",
       " 'drive',\n",
       " 'type',\n",
       " 'paint_color',\n",
       " 'size',\n",
       " 'title_status',\n",
       " 'condition',\n",
       " 'fuel',\n",
       " 'state',\n",
       " 'manufacturer']"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# only want the categorical columns up to manufacturer, description columns already dummied\n",
    "cat_col[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "id": "YeT1wy-lHUj1"
   },
   "outputs": [],
   "source": [
    "def get_dummies(df):\n",
    "  '''\n",
    "  This function takes the train/test dataset and creates dummy variables for the known categorical variables.\n",
    "  '''\n",
    "  # using pandas get dummies function to create dummy variables for each categorical item\n",
    "  dummy_df = pd.get_dummies(df[cat_col[:10]])\n",
    "\n",
    "  # saving dummy feature names to list\n",
    "  dummy_features = list(dummy_df.columns)\n",
    "\n",
    "  # merging original df w/ the new dummy variable columns\n",
    "  new_df = df.merge(dummy_df, left_index=True, right_index=True)\n",
    "\n",
    "  # return/output the new merged df and the dummy feature column names\n",
    "  return new_df, dummy_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "id": "dTvWS98DK0kJ"
   },
   "outputs": [],
   "source": [
    "# using function to create new large df and the list of dummy column names\n",
    "new_large_df, dummy_col = get_dummies(large_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 351
    },
    "id": "B_gr-AXhcKLR",
    "outputId": "f95f68bc-1cd3-420b-a63b-422fa5aee2eb",
    "scrolled": true
   },
   "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>id</th>\n",
       "      <th>price</th>\n",
       "      <th>year</th>\n",
       "      <th>odometer</th>\n",
       "      <th>cylinders</th>\n",
       "      <th>transmission</th>\n",
       "      <th>drive</th>\n",
       "      <th>type</th>\n",
       "      <th>paint_color</th>\n",
       "      <th>size</th>\n",
       "      <th>...</th>\n",
       "      <th>manufacturer_pontiac</th>\n",
       "      <th>manufacturer_porsche</th>\n",
       "      <th>manufacturer_ram</th>\n",
       "      <th>manufacturer_rover</th>\n",
       "      <th>manufacturer_saturn</th>\n",
       "      <th>manufacturer_subaru</th>\n",
       "      <th>manufacturer_tesla</th>\n",
       "      <th>manufacturer_toyota</th>\n",
       "      <th>manufacturer_volkswagen</th>\n",
       "      <th>manufacturer_volvo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7234568651</td>\n",
       "      <td>44989</td>\n",
       "      <td>2015.0</td>\n",
       "      <td>71639.0</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>automatic</td>\n",
       "      <td>4wd</td>\n",
       "      <td>truck</td>\n",
       "      <td>white</td>\n",
       "      <td>full-size</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7229694925</td>\n",
       "      <td>39988</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>108531.0</td>\n",
       "      <td>5.888874</td>\n",
       "      <td>automatic</td>\n",
       "      <td>4wd</td>\n",
       "      <td>other</td>\n",
       "      <td>silver</td>\n",
       "      <td>full-size</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7237581553</td>\n",
       "      <td>25990</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>48931.0</td>\n",
       "      <td>5.888874</td>\n",
       "      <td>automatic</td>\n",
       "      <td>4wd</td>\n",
       "      <td>other</td>\n",
       "      <td>blue</td>\n",
       "      <td>full-size</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7235261085</td>\n",
       "      <td>18990</td>\n",
       "      <td>2011.0</td>\n",
       "      <td>97792.0</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>other</td>\n",
       "      <td>4wd</td>\n",
       "      <td>sedan</td>\n",
       "      <td>white</td>\n",
       "      <td>full-size</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7235871752</td>\n",
       "      <td>39995</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>121365.0</td>\n",
       "      <td>5.888874</td>\n",
       "      <td>automatic</td>\n",
       "      <td>4wd</td>\n",
       "      <td>sedan</td>\n",
       "      <td>white</td>\n",
       "      <td>full-size</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 166 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           id  price    year  odometer  cylinders transmission drive   type  \\\n",
       "0  7234568651  44989  2015.0   71639.0   8.000000    automatic   4wd  truck   \n",
       "1  7229694925  39988  2017.0  108531.0   5.888874    automatic   4wd  other   \n",
       "2  7237581553  25990  2017.0   48931.0   5.888874    automatic   4wd  other   \n",
       "3  7235261085  18990  2011.0   97792.0   8.000000        other   4wd  sedan   \n",
       "4  7235871752  39995  2017.0  121365.0   5.888874    automatic   4wd  sedan   \n",
       "\n",
       "  paint_color       size  ... manufacturer_pontiac manufacturer_porsche  \\\n",
       "0       white  full-size  ...                    0                    0   \n",
       "1      silver  full-size  ...                    0                    0   \n",
       "2        blue  full-size  ...                    0                    0   \n",
       "3       white  full-size  ...                    0                    0   \n",
       "4       white  full-size  ...                    0                    0   \n",
       "\n",
       "  manufacturer_ram manufacturer_rover manufacturer_saturn  \\\n",
       "0                0                  0                   0   \n",
       "1                0                  0                   0   \n",
       "2                0                  0                   0   \n",
       "3                0                  0                   0   \n",
       "4                0                  0                   0   \n",
       "\n",
       "   manufacturer_subaru  manufacturer_tesla  manufacturer_toyota  \\\n",
       "0                    0                   0                    0   \n",
       "1                    0                   0                    0   \n",
       "2                    0                   0                    1   \n",
       "3                    0                   0                    0   \n",
       "4                    0                   0                    0   \n",
       "\n",
       "   manufacturer_volkswagen  manufacturer_volvo  \n",
       "0                        0                   0  \n",
       "1                        0                   0  \n",
       "2                        0                   0  \n",
       "3                        0                   0  \n",
       "4                        0                   0  \n",
       "\n",
       "[5 rows x 166 columns]"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# previewing new data set, it has 166 columns now counting non-dummy features\n",
    "new_large_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "id": "tjBEOKLIlB4R"
   },
   "outputs": [],
   "source": [
    "# Set X and y\n",
    "X = new_large_df.loc[:, num_col + dummy_col]\n",
    "y = new_large_df['price_target']\n",
    "\n",
    "# splitting the data into train and test\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.7, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IKfs8679K2Ds"
   },
   "source": [
    "# Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "uu6lzuqCK4WM",
    "outputId": "e7932b32-e7c0-4a77-b04f-058fa52edb3e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log model Cross Val Score: 0.7738285714285714\n"
     ]
    }
   ],
   "source": [
    "# create logistic regression model\n",
    "log_model = LogisticRegression()\n",
    "\n",
    "# try cross val score\n",
    "score = cross_val_score(log_model, X_train, y_train).mean()\n",
    "\n",
    "# display score\n",
    "\n",
    "print(f'Log model Cross Val Score: {score}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "6cjJiYdepl2D",
    "outputId": "6aabbc72-c004-43b1-affe-a15f886c78a5"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression()"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# create logistic regression model\n",
    "log_model = LogisticRegression()\n",
    "\n",
    "# fit logistic regression model\n",
    "log_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "yTzm9g4vLFTZ",
    "outputId": "6865ced1-cba7-4657-c6d0-c313dd58d5d7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log Model Training Accuracy Score: 0.7603619047619048\n"
     ]
    }
   ],
   "source": [
    "# train accuracy score\n",
    "log_train_accuracy = log_model.score(X_train, y_train)\n",
    "\n",
    "# display training accuracy score\n",
    "print(f'Log Model Training Accuracy Score: {log_train_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OmIYUJGfLLu3",
    "outputId": "c0e10cf2-323b-49d4-cd3e-41ad2e0534bf"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log Model Training Accuracy Score: 0.7597555555555555\n"
     ]
    }
   ],
   "source": [
    "# testing accuracy score\n",
    "log_test_accuracy = log_model.score(X_test, y_test)\n",
    "\n",
    "# display test accuracy score\n",
    "print(f'Log Model Training Accuracy Score: {log_test_accuracy}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uJ_Gg-TTKiLv"
   },
   "source": [
    "# Decision Tree Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "1X3N4PMikSRL",
    "outputId": "11ddff68-f872-46af-eec7-22e2369cfac5"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier()"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# create decision tree model\n",
    "dt_model = DecisionTreeClassifier()\n",
    "\n",
    "# fit decision tree model\n",
    "dt_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "myUFqH3vyko6",
    "outputId": "421ad69d-0ad9-4762-f649-15b0a8781c13"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DT Model Training Accuracy Score: 0.9997142857142857\n"
     ]
    }
   ],
   "source": [
    "# train accuracy score\n",
    "dt_train_accuracy = dt_model.score(X_train, y_train)\n",
    "\n",
    "# display training accuracy score\n",
    "print(f'DT Model Training Accuracy Score: {dt_train_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "TWbs9fu4yqKX",
    "outputId": "1a2709f0-8c48-49f1-d83d-ae95c7aa65ee"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DT Model Test Accuracy Score: 0.8935\n"
     ]
    }
   ],
   "source": [
    "# testing accuracy score\n",
    "dt_test_accuracy = dt_model.score(X_test, y_test)\n",
    "\n",
    "# display test accuracy score\n",
    "print(f'DT Model Test Accuracy Score: {dt_test_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "id": "ZNuTonBrNIOI"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(estimator=DecisionTreeClassifier(),\n",
       "             param_grid={'max_depth': range(1, 20, 2),\n",
       "                         'max_features': [None, 'auto'],\n",
       "                         'min_samples_leaf': [1, 2, 5, 10]})"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# try using grid search to find better parameters\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# create decision tree model\n",
    "dt_model = DecisionTreeClassifier()\n",
    "\n",
    "# define parameters to test\n",
    "parameters = {'max_depth':range(1,20,2),\n",
    "              'min_samples_leaf':[1, 2, 5, 10],\n",
    "              'max_features':[None,'auto']\n",
    "              }\n",
    "\n",
    "# creat grid search to find best parameter\n",
    "clf = GridSearchCV(dt_model, parameters)\n",
    "\n",
    "# fitting our grid search model\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "iyqhECrfzpwK",
    "outputId": "10100a07-7777-4572-8e03-277a2016631e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best DT Model Parameters: {'max_depth': 19, 'max_features': None, 'min_samples_leaf': 1}\n"
     ]
    }
   ],
   "source": [
    "# storing best parameter\n",
    "best_param = clf.best_params_\n",
    "\n",
    "# display best parameter\n",
    "print(f'Best DT Model Parameters: {best_param}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "QSMXwZRXb625",
    "outputId": "49936853-87f1-4f28-e6ec-1bfe6920d7a3"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(max_depth=19)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# use the parameters from above to create decision tree model\n",
    "# create decision tree model\n",
    "dt_model = DecisionTreeClassifier(max_depth=19,\n",
    "                                  min_samples_leaf=1)\n",
    "\n",
    "# fit decision tree model\n",
    "dt_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "L04A5biV0apX",
    "outputId": "df4a53da-8b19-4930-d979-0834d5dd55b6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DT Model Training Accuracy Score: 0.9566428571428571\n"
     ]
    }
   ],
   "source": [
    "# train accuracy score\n",
    "dt_train_accuracy = dt_model.score(X_train, y_train)\n",
    "\n",
    "# display training accuracy score\n",
    "print(f'DT Model Training Accuracy Score: {dt_train_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "yqO-ep5Q0apY",
    "outputId": "b9be1809-ce27-4d3c-e926-85c172c28aea"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DT Model Test Accuracy Score: 0.9019\n"
     ]
    }
   ],
   "source": [
    "# testing accuracy score\n",
    "dt_test_accuracy = dt_model.score(X_test, y_test)\n",
    "\n",
    "# display test accuracy score\n",
    "print(f'DT Model Test Accuracy Score: {dt_test_accuracy}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cjyeUua8O-5C"
   },
   "source": [
    "# KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "id": "w8eRJrfBtPpp"
   },
   "outputs": [],
   "source": [
    "# testing various K's to find the best non-overfitted value\n",
    "k_value = []\n",
    "training_accuracy = []\n",
    "testing_accuracy = []\n",
    "\n",
    "# use for-loop and cross_val_score to find the best k value\n",
    "for k in range(1, 11):\n",
    "    # add k to k_value list\n",
    "    k_value.append(k)\n",
    "\n",
    "    # create KNN model\n",
    "    knn_model = KNeighborsClassifier(n_neighbors=k)\n",
    "\n",
    "    # fit model with our X and y\n",
    "    score = cross_validate(knn_model, X_train, y_train, n_jobs=-1, return_train_score=True)\n",
    "\n",
    "    # appending average accuracy per K\n",
    "    training_accuracy.append(score['train_score'].mean())\n",
    "\n",
    "    # appending average accuracy per K\n",
    "    testing_accuracy.append(score['test_score'].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "id": "3VpiYwmavPVq"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot knn_cv_score vs k_value to find best k_value\n",
    "plt.figure(figsize=(12, 8))\n",
    "plt.plot(k_value, training_accuracy)\n",
    "plt.plot(k_value, testing_accuracy)\n",
    "plt.xlabel('K Value')\n",
    "plt.ylabel('10-Fold Accuracy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtcAAAHgCAYAAABuGUHVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3ic1Zn+8fuoy5Ysuai4yb03bORuWmiGEMwmQGwSejAOkJ7dsNn9bfpu2maxIYEQIKRiCCGhd0KLbWwZd7lXyU2uslzUz++PM7JGsmSPpJl5p3w/1zWXNTPvWI/wi3TrzHmfx1hrBQAAAKDjErwuAAAAAIgVhGsAAAAgSAjXAAAAQJAQrgEAAIAgIVwDAAAAQUK4BgAAAIIkyesCgqlHjx62f//+XpcBAACAGLZ8+fKD1tqclp6LqXDdv39/FRUVeV0GAAAAYpgxZmdrz7EtBAAAAAgSwjUAAAAQJIRrAAAAIEgI1wAAAECQEK4BAACAICFcAwAAAEFCuAYAAACChHANAAAABAnhGgAAAAgSwjUAAAAQJIRrAAAAIEgI1wAAAECQEK4BAACAICFcAwAAAEFCuAYAAACChHANAAAABAnhGgC8UFvtdQUAgBBI8roAAIgb1Sek9S9Jq56Str0rde4h5Y6QckdJeSOl3JFSznApNcPrSgEA7US4BoBQqq+TdnwgrVooFb8g1ZyQsgukqfdKlUel/cXSx7+Tak42vqZrfxe0c0f6QvcoqfsgKTHZsy8DABAYwjUAhELZBmn1Qmn1M9Kx3VJqF2nMZ6Rxc6S+U6QEv1159fXS0R0uaJf5bvuLpU2vS7bOHZOYIvUY6gvdI6S8Ue7jrD6SMZ58iQCAMxGuASBYThyU1jzrtn3sXSmZRGnwZdIVP5SGXSUlp7f8uoQEqdtAdxtxTePjNZXSoc2+0L3O/blzkbTmmcZjUrv4tpaMbAzcuSOkTt1C+7UCAFpEuAaAjqiplDa95rZ9bHlTqq+Veo6TZv5YGv0ZKSO3/X93cpqUP8bd/J06KpWtb7rKve45aflvG4/J7HnmKnfOsNYDPgAgKAjXANBW1kolH7kV6rV/k6rKXZideq80drbbJx1K6dlSv6nu5l9Txd7GVe6y9dL+ddLSD6W6KneM8a2Qn17l9l1M2W2AlJAY2poBIE4QrgEgUIe3uT3Uq56SjuyQkjtJI66Vxs2WBlzobUA1RurSy92GXNb4eF2tdGS7C9plxe7P/euk9S9Ksu6YpDS3qn26a4kvdGfms58bANrIWGu9riFoCgsLbVFRkddlAIglp45I6/7utn2ULJFkXJAeN0ca8anobZtXfVI6sKFxe0lD+D6+v/GY9K4uZOeOaOxakjtCSuviXd0AEAGMMcuttYUtPcfKNQA0V1cjbXnLrVBvfM1tq+gxTLrsu9KYG6Ws3l5X2HEpnaTeE9zN34lDjXu5G/Zzr1ooVVc0HpPV169NoO/WY6iUlBLerwEAIhDhGgAkt2d5zwpp9dOu48fJg1KnHlLh7W7bR8/z4mOLROfu0oAL3K2BtdLRXb5Vbl/XkrL10tZ3pPoad0xCktR9yJmr3Nn9mrYdBIAYR7gGEN/KS337qBdKBze6ftLDrnbbPgZfyuAWyf1S0bWfuw2b2fh4bbV0aEvTVe7dRa5zSYOUDDd10r9rSd4oN50SAGIQ4RpA/Kk67i7oW/WUtP19SdYNdrnmAWnUdW6vMc4tKcWtUjfvjlJV4Ybo+Hct2fiKtOIPjcd0zjmza0nucCmlc3i/BgAIMsJ1R9XVuEERvO0JRLb6Omn7e26Fev2Lbtx41/7SxfdLY290LeoQHKmZUt+J7tbAWul4WdNV7rJiafmTfqPffSvkp7uW+G7dB0uJ/LgCEB34btUR1kovfkWqrZRm/coNfAAQWfYXuxXqNX9xfaDTsqSxn3X7qPtOjo991JHAGCkzz90GXdL4eH29axVYVty4yl1W7AbzNBn9Pqxpm8C8kVKX3vz7AYg4hOuO6jFEeuu7bt/m7KfcxUAAvHW8rHEM+b7V7mK7wZe7qYlDZ/KLcCRJSJC6D3K3EZ9qfLymUjq4ya9N4Hppx4fugtMGqVl+F1CObOxgwrYeAB6iz3UwrPub9NzdbnjD556VegwO/LUV+6Si37oJaeNmh67GeGKtdGw3q1rxpuaU29e76mnXRs/WSb3GuwsTR3+GC+hixakjjfu59/ttMakqbzwms9eZXUtyhvNLFYCgOVufa8J1sJQslZ6a436gz/6z1G/a2Y8v2yAtftB1Kairdo+d93npkz+XktNDX28sqqt1XQo+fMD94M0fK037srtAjY4Psam+3g12WfWUG/RSdcz9UjX2RjeGPHe41xUiHKyVju1ptp97nXRgU7PR74OarXKPcvvuGf0OoI0I1+FyeLv0pxukozulWb90P+D9Weve1lz0oLT5dSkpXRr/eWnyPGnNM9J7P5Hyxkg3/s69RRps9fVuZe+f890PnJk/kfpNDf7nCbeaU9KKP0qLFrhevDnDpVH/Iq39q3tbuUsfaco8acKtTJaLFYe2ugsTVy90/+bJnaWRs6Rxn5X6X0BYglNX60bW+69ylxW779WnR7+nu9HvDW0CG1oGZuTxzheAVhGuw+nUEenpm6UdH0gXf1u66N9cl4L1z7tQvWeFG0wx+W6p8M6me7Q3vyk9d5c7/rqHpRHXBKemuloXND/8hRt33LW/C9rlJdKku6RLvxOdI5wry6Vlj0tLHpZOlEl9Jkozvu721CYkuK9xy5vuv/uOD6TULtKEW6QpX5Sy+nhdPdrq5GG3BWvVQql0qSQjDbzYN4b8Glq4IXDVJ6QDG5uucpetbzb6vVvTKZR5o9wv7vyCDkCE6/CrrZZe/LJ7q3rIFS7QHt3l2klNvc/trW5t68fRXdIzt7gQPv0r0if+q/0tqGoqpZV/dCvVR3e5HxAXfEMaeZ3rcPL296Wlj0rZfaVPzZcGfaL9X3M4HS+TlvzKBeuqY9KgS6ULvi71m976StOeFdKih1w4k6TRn3b/Fr3OC1/daLvaavcL0qqnpE2vuy1UOSOk8+ZIY25w1zkAwXLiUNPe3A0dTKqPNx6TVXBm15LuQxj9DsQZwrUXrJXe/5n0jx9JBVOlaV+Shl4VWD/s2irptX+Xih6X+s2Qrn/Cta8KVOUxqegJafEvG1d0L/iGNOTKMz//riXS8/dJhza7LSpX/EhKz27b1xouR3ZI/1zgtoDUVbu91NO/2raAfLRE+ugRafnvpOoKt4Vg2pelwZfRqzxSWCvt/tht+VjzrHTqsBs4MuYG94tp/ljerkf4NLzL59+1pKzYbTmrr3XHNIx+b961JKuA7ytAjCJce+nU0faH1VVPuz7aaV2k638r9Z9+9uNPHHLBcemv3ZaJgZe4UN1/xtnDSE2l9N6PXXDtnCNd8wtp+CfbV3Mo7F8nffh/0trn3F7acXPcqn5H9qVXlruA/dEjrrNIj2HS1Htd/2M6CnjjaIlrs7ZqoftlLzHVnYfj5ri+yFyUikhSW+3O0yar3MXuXcIGDaPfG7qWNIRvOtcAUY9wHc32F0vP3OwuwLnsu24FvHlQLt8tLX6ocdLZiE+5vce9J7Ttc+1Z4Vax9691rcuu+qm3PwR2LZE++IW7+DMlQzr/NheAg7kVoK7GbRVZtEDat8b9cjHpbmninVKnbsH7PGhZ5TFp/QsuUO/4wD3Wb7r7JWfkrMh9FwVoTeUxtxXQfwrl/nXuHZgGnXPPXOXOYfQ7EE0I19Gu8pj0/L0uhAy/RrruV27K3KGt0j8fkFY+Jdl6151k+lc71n6sttr9ne/91K2YX/VTF7TD9Ta8te7Czg9/Ie1a7C4qmvJFaeIXQht2rZW2v+8uftzypq+Ty+ekKfeEpnNLPKurlba/6xtD/pJUe8qNHh83x53DXft7XSEQXKdHvzd0LVnv29u9wZ3/ktzo9/5ndi3pNojR70AEIlzHAmvdRXxv/D+paz+p5zip+HkpIdl1wJj2Jfd4sJStd4F+93K3V/yaX4T24rG6Wqn47277x/61rn3etC9JE24O/2pO2Xr3TsDqZ9zK9vBPun3ZBZPDW0es2bfWN4b8Wen4Pikt211YOm6Ouy6AfdSIN/V17lqS5l1LDm1xCyaS2x6VM7Rpb+7cEQzJAjxGuI4lOxdLz94uVR13Wxem3NO2ix3bor7Otbl754dSYop0xQ9ckA/mN/SaSmnln9y2jCM73N7nGV91F695vce2Yr/rplL0uGux2Gei6zAy4lP0UQ5UxX5pzV/cKvX+Ne7CryFXugsTh14pJaV6XSEQeWoqpYMbm+7n3l8sVexpPCYtq3GF2z90M/odCAvCdaypOi7JSqmZ4fl8h7ZKL3xZ2vmh6yv8qfkdf+u+oaPJkl+53rK9z3f7xIddHXlX11efkFb+2XVfObLdfe1T7pHO+1x09gcPteqTvjHkT0lb33ErcL3PdyvUoz7dtLc7gMCdOuIXuNe3Pvq9+X7uHsO4UBsIMsI1Oq6+Xlr+W+nN77gR75d+R5o0t+1B+PgB6aOHpaWPuR8IAy+RZnxNGnBh5L/FWV8nbXjZbRkp+chta5h4p/vvkJnvdXXeqq+Xdv7Ttc9b97xrc9ilj5uYOHa2e1sbQPBZ6zoeNe9acmCja1kqSSbRXTvi35s7dySj34EOIFwjeMpLpRe/6i766ztZuvahwILTkZ3uYsEVf3B9vEde6y6+bGtHk0hRstR9PetfdNtXxtzgtozkjfS6svA6uNk3hvxp1ws4JcMNKRo323X9iLR3IYB4UVcrHd7abJV7ndt+5z/6PXe4C9y5IxpbBmbkRv5iB+AxwjWCy1oXpl79llRzSrr4fnfBX0tXtO8vdt1H1jwrmQQXuqZ/ReoxJPx1h8LhbW5f+oo/ujaIgy51F2IOvDh2fzidPCyt/asL1buL3L/rwEvcto/hn5RSOnldIYDWVJ9wrQIb2gQ2bC05UdZ4THq3xq4lp7eYjAjfVkQgChCuERoV+6VXvulaBPYcJ836pZQ/xj1XstT1qN70qpTcubFHdVZvT0sOmZOH3YWPHz3qfkjljZGm3ef2GMfCWOTaKmnzGy5Qb3pdqq9xK1wNY8jjfVsMEO1OHDyza0nz0e/ZBWd2LWH0O+IU4RqhVfy89PI33ZCESXOlvavc/tv0rtLkee6xeBnIUlvlWvgtfsitDmX2kibf7X65iLaBKNZKpUVuH/Xav7qLqTLy/MaQj/G6QgChVF8vle86c5X70Ga/0e/J7p1I/97cuSOlrL5sC0NMI1wj9E4ell7/tusQ0aW32398/q3xO3HMWmnL267F4Pb33F7kCbe4XzaC2Y88FI7sdL8grHrK7dlMSnPDi8bNcdtdGGgBxLeG0e/+q9z7i10Qb5CScWabwNxRdAtCzCBcI3zKS91oX94mbLR3tVvJXvtXF7pHznL7siPpYs7KcvcOxKqF7l0HSeo3w61Qj5zlpnUCwNk0jH73781dts6969UgI69pm8DchtHvXKuB6EK4BiJB+W7po0ek5U9KVcdcN41pX3JDVbx4+7Su1vWhXr3QtRisrZS6D3aBesyNkb/CDiDyWetmGTTvWnJgg/ueI0kyUrcBzUL3KKnbQN4pQ8QiXAORpPKYa0m45GHXvq77EHex57jZUnJ6aD+3tdK+NW6Fes1f3MWX6V2l0de7z9/7/NjtcgIgcjSMfvfvzb2/2G1FO2P0+6img3G69OL7FDznWbg2xsyUNF9SoqTHrLU/bvb85yR9y3f3uKQvWmtXBfLalhCuEVXqaqXiv7t+2XtXSp16SJPukiZ+QercI7if69jexjHkZevcRUhDr3T7qIdcwTYeAJGh5pR0cFPjlpL9xW7Fu8XR736r3Lkjou+icUQ1T8K1MSZR0iZJl0sqlbRM0hxrbbHfMdMkrbfWHjHGXCXpu9bayYG8tiWEa0Qla90+50UPSptecxcQjpvjVrM70g+8+oTb7rHqKWnbu241qM9Et0I96tPx08EFQPQ7ebhxW8np/dzrm45+79L7zK4lOcOkpFTv6kbMOlu4DuVmpkmStlhrt/mKWChplqTTAdlau8jv+CWS+gT6WiBmGCP1n+FuBzZKi38prfyzGzc/7GrXeaXftMDeBq2vl3Z84Fao17/getRmFUgXfMONIe8xOPRfDwAEW6duUv/p7tagYfR7864l299rYfT7yMbAnTtC6jqAVoEImVCG696SSvzul0qafJbj75T0ajtfC8SGnGHStQukT/w/adlvpKW/kTa+IvWa4C5+HHFtyxf4HNjoG0P+jHSsVErJlEb9i1sBL5jKDxEAsccYKauPuw29ovHxuhrp0Namq9x7V7mOSA2j35M7uS4lTaZQjmT0O4IilOG6pbOzxT0oxphL5ML1jHa8dq6kuZJUUFDQ9iqBSJSRI13ybWn6V922jsW/lJ693a1CT/miNOFmN7CmYQz5no/dCs2gT0iXf8+teNPaCkA8SkyWcoe7mz7d+Hj1Calsg98q9zpp8+vSyj82HtOp+5m9uXOHM/odbRLKPddT5fZQX+m7/++SZK39n2bHjZX0N0lXWWs3teW1zbHnGjGrvt6Nkl/0oLRrsVuZrj3lpqTlj3Er1KOvlzLzvK4UAKLL8QONq9z++7lrTjQek11wZteSHkNckEdc8uqCxiS5ixIvlbRb7qLEm6y16/yOKZD0jqRb/PdfB/LalhCuERdKi6Si30qdurp91Pmjva4IAGJLfb10dKfvIkq/riVnjH4f6ruA0q9rSXYBW0vigCcXNFpra40x90l6Xa6d3hPW2nXGmHm+5x+R9F+Sukv6lXEnYq21trC114aqViCq9Cl0NwBAaCQkuME23QZIw69ufLy2Sjq4uWnoLvlIWvts4zEpmX6Be2TjNhM6NMUNhsgAAAB0RGV5437uhlXuM0a/5zdtE5g7gtHvUcyrVnwAAACxLy1LKpjsbg2slSr2NdvLvU5a9liz0e8Dm4buvFGuVSCj36MW/3IAAADBZozUpae7Db608fH6Ounw9qZdS8qKXdvVJqPfhzXtWpI3UsrsyX7uKEC4BgAACJeERDfQq8dgaeSsxsdrTrmZBWXFvsC93k3XXfVU4zFp2Wf25mb0e8QhXAMAAHgtOV3qdZ67+fMf/d4Qulc/I1UdazymS5+mXUvyRrpOJox+9wThGgAAIFK1Nvq9vLTZfu5it9JdX+OOMYlS98FN2wTmjZSy+zO1N8QI1wAAANHEGCm7r7sNvbLx8dOj3/26luxZIa37W+MxDaPfm4TuUW70O4KCcA0AABAL/Ee/j/5M4+NVx6UDG5p2Ldn0urTCf/R7jzN7c+cMl1Izwv91RDnCNQAAQCxLzWh5ANnxA2d2Lfn4D81Gv/fz61riC93dBzP6/SwI1wAAAPEoI0fKuFgaeHHjY6dHv/vt5S4rdivdts4d0zD6Pc9vlTt3hJTVl1aBIlwDAACgQZPR759sfPz06PfixlXuXUukNX9pPCa1i2+F2683d+7IuBv9TrgGAADA2SWlSvmj3c1fZblfq0DfKve6v0vLn2w8JiO/2X7ukW4/d3J6WL+EcCFcAwAAoH3SsqSCKe7W4PTo93VNt5b4j343CW7Mu39v7tyRbhx8QqI3X0uQEK4BAAAQPE1Gv1/W+Hh9nXR4W9OuJWXrpQ0vN45+T0pzo9/9V7lzo2v0O+EaAAAAoZeQKPUY4m5njH7f0HSVe+s/zhz9njeqceR7w0WUaVnh/zrOgXANAAAA7ySnS73Gu5u/k4fP7Fqy+ummo98v+Q/pon8Lb73nQLgGAABA5OnUTeo/w90aWCuVlzT25vbf6x0hCNcAAACIDsZI2QXu5j/6PYIkeF0AAAAAECsI1wAAAECQEK4BAACAICFcAwAAAEFCuAYAAACChHANAAAABAnhGgAAAAgSwjUAAAAQJIRrAAAAIEgI1wAAAECQEK4BAACAICFcAwAAAEFCuO4ga61OVdd5XQYAAAAiQJLXBUQza63u/sNySdKvbz5fxhiPKwIAAICXWLnuAGOMJvbvpjeK9+uZohKvywEAAIDHCNcddOeMAZo2qLu+92Kxth884XU5AAAA8BDhuoMSEoz+98ZxSkow+urTK1VTV+91SQAAAPAI4ToIemal678/PUarSo7qoXe2eF0OAAAAPEK4DpJrxvbSpyf01oPvbNbynUe8LgcAAAAeIFwH0feuHaVe2en62tMrdbyq1utyAAAAEGaE6yDKTEvW/332PJUeOanvv7jO63IAAAAQZoTrIJvYv5vuuXiwnikq1Wtr93pdDgAAAMKIcB0CX7lsiMb2ydL9z63R/mOVXpcDAACAMCFch0ByYoIe+Ox5qqqp1zf/skr19dbrkgAAABAGhOsQGZiTof+8ZoQ+2HxQVy/4QO9s2C9rCdkAAACxjHAdQjdNKtCCOeN1qqZOdzxZpBt/vVjLdhz2uiwAAACECOE6hIwxunZcL7319Yv0w+tGa+ehk7rhkcW648llKt5zzOvyAAAAEGQmlrYqFBYW2qKiIq/LaNWp6jo9uWiHHn53iyqqanXtuF76+uVD1a97Z69LAwAAQICMMcuttYUtPke4Dr/ykzX69ftb9cQ/t6u2zmr2pL766mVD1SMj1evSAAAAcA5nC9dsC/FAVqdk/dvM4Xr/Xy/RnEkFWri0RDf9ZolOVdd5XRoAAAA6gHDtodwuafrBdaP1xG0TtbnsuL7zwlqvSwIAAEAHEK4jwIVDc3Svb6rj31aUel0OAAAA2olwHSG+etkQTerfTf/xt7XaUnbc63IAAADQDoTrCJGUmKAFc8YrLTlR9/35Y1XWsP8aAAAg2hCuI0h+Vpp+ceM4bdhXoe+9WOx1OQAAAGgjwnWEuXhYruZdNEhPLd2lF1bt8bocAAAAtAHhOgJ944qhKuzXVf/+19XafvCE1+UAAAAgQITrCJTs23+dnJSge/7E/msAAIBoQbiOUL2y0/WLG8dp/d5j+uHL7L8GAACIBoTrCPaJ4Xmae+FA/XHJLr20mv3XAAAAkY5wHeH+9cphGl+Qrfv/ukY7D7H/GgAAIJIRriNccmKCHpwzXokJRvf++WNV1bL/GgAAIFIRrqNAn66d9LPrx2rt7mP6n1c2eF0OAAAAWkG4jhJXjMrXHdMH6MlFO/Ta2r1elwMAAIAWEK6jyP1XDde4Pln612dXq+TwSa/LAQAAQDOE6yiSkpSgh26aIEn69MOL9Oj7W3W8qtbjqgAAANCAcB1l+nbrpD/eOVnD8jL1369s0PQfv6NfvLFRh09Ue10aAABA3DPWWq9rCJrCwkJbVFTkdRlhs6rkqH717ha9vm6/0pITNHtigeZeOFC9stO9Lg0AACBmGWOWW2sLW3yOcB39tpRV6OF3t+n5lbslSdeN7615Fw3S4NwMjysDAACIPYTrOLH76Cn95v1tWrhsl6pq63XlyHzdc8kgje2T7XVpAAAAMYNwHWcOHa/Sk4t26MlFO1RRWasZg3vonosHaeqg7jLGeF0eAABAVCNcx6mKyhr96aNdevzD7TpQUaVxfbN1z8WDdPmIPCUkELIBAADag3Ad5ypr6vTs8lL9+v2tKjl8SkNyM/TNK4fpylH5XpcGAAAQdc4WrmnFFwfSkhP1+Sn99I9vXKz5s8+TMdK8Py7XK2uY9AgAABBMhOs4kpSYoFnn9dbz987Q+L7Z+urClVqy7ZDXZQEAAMQMwnUcSk9J1OO3TlRB906663dFWr/3mNclAQAAxATCdZzq2jlFv7tjkjqnJunWJ5aq9MhJr0sCAACIeoTrONY7O12/u2OSTtXU6ZYnljJCHQAAoIMI13FuWH6mHrulUKVHTumOJ5fpZHWt1yUBAABELcI1NHlgdy2YPV6rS4/qvj+vUE1dvdclAQAARCXCNSRJM0fn6/uzRuudDWX69nNrFEv9zwEAAMIlyesCEDk+P6WfyiqqtODtzcrtkqp/vXK41yUBAABEFcI1mvjaZUN0oKJSv/zHVuVkpOq26QO8LgkAACBqhHRbiDFmpjFmozFmizHm/haeH26MWWyMqTLGfLPZczuMMWuMMSuNMcw0DxNjjH4wa7QuG5Gn771UrJdXM8URAAAgUCEL18aYREm/lHSVpJGS5hhjRjY77LCkL0v6eSt/zSXW2vNam92O0EhKTNBDN43XhIKu+trTK7Vo60GvSwIAAIgKoVy5niRpi7V2m7W2WtJCSbP8D7DWlllrl0mqCWEdaIe05EQ9fmuh+nXvpLt/v1zFe5jiCAAAcC6hDNe9JZX43S/1PRYoK+kNY8xyY8zcoFaGgGR38pvi+NulKjnMFEcAAICzCWW4Ni081pb+btOttRPktpXca4y5sMVPYsxcY0yRMabowIED7akTZ9ErO12/v3OSqmrqdOsTS3XoeJXXJQEAAESsUIbrUkl9/e73kbQn0Bdba/f4/iyT9De5bSYtHfeotbbQWluYk5PTgXLRmqF5mXr8tonaffSU7vhdEVMcAQAAWhHKcL1M0hBjzABjTIqk2ZJeCOSFxpjOxpjMho8lXSFpbcgqxTlN7N9NC+aM15rSo7rnTx8zxREAAKAFIQvX1tpaSfdJel3SeknPWGvXGWPmGWPmSZIxJt8YUyrp65L+0xhTaozpIilP0ofGmFWSlkp62Vr7WqhqRWCuHJWvH143Ru9uPKD7/8oURwAAgOZCOkTGWvuKpFeaPfaI38f75LaLNHdM0rhQ1ob2uWlygcoqKvXAW26K47dmMsURAACgARMa0WZfuXSI9h+r0sPvblVuZqpuZ4ojAACAJMI12sEYox9eN1qHjlfp+y8Vq0dGqj41rpfXZQEAAHgupOPPEbsSE4wWzBmvwn5d9Y1nVmnRFqY4AgAAEK7RbmnJiXrslonq36OT5v5hudbtKfe6JAAAAE8RrtEhWZ2S9bs7JikzLUm3/XYZUxwBAEBcI1yjw3pmpev3d0xSdW29bnliqVaWHKVNHwAAiEuEawTFkLxMPXFbofYfq9R1v/ynpv34HX33hXX6aNsh1dUTtAEAQHwwsbTCWFhYaIuKirwuI66Vn6zRW+v369W1+/T+5gOqrq1Xj4wUXT4yXzNH52vaoO5KTuR3OmPuW9cAACAASURBVAAAEL2MMcuttYUtPke4Rqgcr6rVuxvL9OraffrHhjKdrK5Tl7QkXTYyTzNH5evCoTlKS070ukwAAIA2IVzDc5U1dfpg80G9unav3irer2OVteqUkqhLhudq5qh8XTI8VxmptF0HAACR72zhmjSDsEhLTtTlI/N0+cg81dTVa/HWQ3p17T69WbxPL6/eq5SkBF04JEczR+fr8hF5yuqU7HXJAAAAbcbKNTxVV29VtOOwXlu3T6+v3ac95ZVKSjCaOqi7Zo7O1xUj85WTmep1mQAAAKexLQRRwVqr1aXlenXtPr22dq92HDopY6SJ/bpp5uh8XTk6X72z070uEwAAxDnCNaKOtVYb91fo1TX79Nrafdq4v0KSNK5PlmaO7qmZo/M1oEdnj6sEAADxiHCNqLftwPHTW0dWlbox68PzMzVztGvxNywvU8YYj6sEAADxgHCNmLL76Cm9ttYF7WU7D8taaUCPzrpyVL6uGp2vsX2yCNoAACBkCNeIWWUVlXqzeL9eW7tPi7a6aZC9stJ09Zie+splQ5SZRtcRAAAQXIRrxIWjJ6v11voyvbZ2r/6x8YCG5WXqyTsmKjczzevSAABADDlbuGYONWJGdqcUXX9+Hz1260Q9fmuhth88oc88vEjbDhz3ujQAABAnCNeISRcPy9VTc6foRFWdrn9ksVaVHPW6JAAAEAcI14hZ5/XN1rPzpqpTSqJmP7pE724s87okAAAQ4wjXiGkDczL03D3TNKBHZ33hd0V67uNSr0sCAAAxjHCNmJebmaan756iSQO66evPrNIj721VLF3ICwAAIgfhGnEhMy1Zv719oq4Z21M/fnWDfvDSetXXE7ABAEBwJXldABAuqUmJWjB7vHIyU/XEP7frwPEq/fyGsUpNSvS6NAAAECMI14grCQlG/3XNSOVmpuknr23Q4RNVeuTz5zNsBgAABAXbQhB3jDH64sWD9PMbxmnJtsOa/egSlVVUel0WAACIAYRrxC03cKZQ2w64YTPbD57wuiQAABDlCNeIa5f4D5t5eBHDZgAAQIcQrhH3GobNpKckas5vlui9TQe8LgkAAEQpwjUg37CZL05Tv+6ddeeTy/S3FQybAQAAbUe4Bnxyu7hhMxP7d9PXnl6lR9/f6nVJAAAgyhCuAT9d0pL15B0T9cmxPfXfr2zQD14qZtgMAAAIGH2ugWZSkxL14OzxyslI1eMfbteBiir9/IZxSknid1EAAHB2hGugBQkJRt/51EjldknVT1/bqMMnqvXIzecrI5X/ZQAAQOtYigNaYYzRPRcP1s+uH6vF2w5p9qOLdaCiyuuyAABABCNcA+dwQ2FfPXZLobaWuWEzOxg2AwAAWkG4BgJwyfBc/fmuyaqorNFnHl6k1aUMmwEAAGciXAMBGl/QVc9+cZrSkhM1+9Elep9hMwAAoBnCNdAGg3Iy9Nw901TQrZPueHKZ/r5it9clAQCACEK4Btoor0uanpk3VYX9u+qrT6/Ub97f5nVJAAAgQgQcro0xnUNZCBBNuqQl63d3TNInx/TUj15Zrx8ybAYAACiAcG2MmWaMKZa03nd/nDHmVyGvDIhwqUmJWjBnvG6d2k+PfbhdX3tmpapr670uCwAAeCiQiRj/J+lKSS9IkrV2lTHmwpBWBUSJxASj7147Srld0vSz192wmYc/z7AZAADiVUDbQqy1Jc0eqgtBLUBUMsbo3ksG66fXj9WirYc059ElDJsBACBOBRKuS4wx0yRZY0yKMeab8m0RAdDoxsK++s0t52tzWYWuf2SRdh5i2AwAAPEmkHA9T9K9knpLKpV0nqR7QlkUEK0+MTxPf75rispPuWEza0rLvS4JAACEUSDhepi19nPW2jxrba619vOSRoS6MCBaTSjoqmfnTVNqUqJmP7pYH2xm2AwAAPEikHD9YICPAfAZnOuGzfT1DZt5fiXDZgAAiAettjQwxkyVNE1SjjHm635PdZGUGOrCgGiX1yVNT989VXN/X6SvLFypAxVV+sIFA70uCwAAhNDZVq5TJGXIBfBMv9sxSdeHvjQg+mWlu2EzV43O1w9fXq8fvcywGQAAYlmrK9fW2vckvWeMedJauzOMNQExJS05UQ/dNEHfe3GdfvPBdh2oqNJPrx+nlKSAB6QCAIAoEciki5PGmJ9JGiUpreFBa+0nQlYVEGMSE4y+d+0o5fmGzRxi2AwAADEpkKWzP0naIGmApO9J2iFpWQhrAmLS6WEzn2kcNnPwOMNmAACIJYGE6+7W2scl1Vhr37PW3iFpSojrAmLWjRP76tGb3bCZzzzMsBkAAGJJIOG6xvfnXmPMJ40x4yX1CWFNQMy7dESe/vSFxmEza3czbAYAgFgQSLj+oTEmS9I3JH1T0mOSvhbSqoA4cH6/rnp23lSlJiXqs79erA83H/S6JAAA0EFnDdfGmERJQ6y15dbatdbaS6y151trXwhTfUBMG5ybqb9+0Q2buf3JpQybAQAgyp01XFtr6yRdG6ZagLiUn+WGzYwv6KqvLFypxz7Y5nVJAACgnQLZFrLIGPOQMeYCY8yEhlvIKwPiSFZ6sn5/xyTNHOWGzfzPK+sZNgMAQBQKpMnuNN+f3/d7zEqizzUQRGnJifrl5yboOy+s1a/f36ayiir99PqxSk5k2AwAANHinOHaWntJOAoB4IbN/GDWaOV3SdPP39jkhs18boI6M2wGAICowJIYEGGMMbrvE0P0k8+M0YebD2jObxg2AwBAtCBcAxHqsxML9OjNhdq4r0LXP7xIuw6d9LokAABwDudqxZdgjJl2tmMAhM5lI/P057sm6+ipGn2aYTMAAES8c7Xiq5f0v2GqBUALzu/XTc/Om6qURKPZjy7RP7cwbAYAgEgVyLaQN4wxnzHGmJBXA6BFg3Mz9dw909U7O123/XapXli1x+uSAABAC4y1Z++la4ypkNRZUp2kU5KMJGut7RL68tqmsLDQFhUVeV0GEDLlJ2t01++LtHTHYX1yTE8V9u+q8QVdNbJnF6UkcQkFAADhYIxZbq0tbOm5QFrxZQa/JADtkdUpWb+/c5J+9PJ6vbV+v15es1eSlJKUoDG9szS+b7Ym9Ouq8QXZ6pmV7nG1AADEn3OuXEuSMeZaSRf67r5rrX0ppFW1EyvXiDd7y09pxa6jWrHriD7edVRrdperurZektQzK03jC7I1ocCF7VG9spSWnOhxxQAARL8OrVwbY34saaKkP/ke+ooxZoa19v4g1gigHXpmpavnmHRdPaanJKm6tl7Fe49pxa4jWrHrqD7edUSvrNknSUpONBrZK0sTCrI1vqCrxvfNVp+u6eJyCgAAgieQPderJZ3n6xwiY0yipBXW2rFhqK9NWLkGzlRWUelb3XZhe3XpUVXWuNXtnMzUxq0kfbM1tk+20lNY3QYA4Gw6tHLtky3psO/jrKBUBSAscjPTdOWofF05Kl+SVFNXr437Kk5vJVmx64jeKN4vyY1fH9Ez8/RWkgkFXVXQrROr2wAABCiQlevZkn4i6R9ynUIulPTv1tqFoS+vbVi5Btrn0PEqrSxpXN1eVXJUJ6rrJEndO6dovN9WknF9s9U5NdDfywEAiD3tXrk2xiRIqpc0RW7ftZH0LWvtvqBXCcAz3TNSdemIPF06Ik+SVFdvtWl/xemwvWLXEb21vkySlGCkoXmZp7eSTOjXVQO6d1ZCAqvbAAAEsnL9vrX2wrMeFCFYuQZC5+jJ6iar2ytLjqqislaSlJWe7Fa3+7rtJOcVZKtLWrLHFQMAEBod3XP9pjHmm5KelnSi4UFr7eHWXwIg1mR3StHFw3J18bBcSVJ9vdXWA8f9VreP6r1Nm2StZIw0OCejce92v64anJPB6jYAIOYFsnK9vYWHrbV2YGhKaj9WrgFvHaus0eqS8tNbSVaUHNXRkzWSpMzUJJ1XkK3xfbM13relJLtTiscVAwDQdmdbuT5ruPbtub7BWvt0qIoLJsI1EFmstdp+8EST1e0N+46p3vdtZ2CPzu5CSV9nkqF5GUpKZIw7ACCytTtc+17MnmsAQXOiqlarS8tPh+0Vu47o0IlqSVKnlESN65N9OmyfV5CtHhmpHlcMAEBT7LkGEDE6pyZp6qDumjqouyS3ul1y+FSTrSSPvr9Ntb7l7YJunU5PlZxQ0FXDe2YqmdVtAECECumea2PMTEnzJSVKesxa++Nmzw+X9FtJEyT9h7X254G+tiWsXAOx4VR1ndbuKdfHOxvHuJdVVEmS0pITNLZ39une2xMKspXbJc3jigEA8aRD20I68EkTJW2SdLmkUknLJM2x1hb7HZMrqZ+k6yQdaQjXgby2JYRrIDZZa7WnvPJ02F5RckTrdh9TdZ0b4947O/30VpLxBdka2auLUpMY4w4ACI12bQsxxvybtfanvo9vsNb+xe+5/7bWfvscn3eSpC3W2m2+1yyUNEvS6YBsrS2TVGaM+WRbXwsgfhhj1Ds7Xb2z0/Wpcb0kSZU1dSree8wF7pKj+njnEb20eq8kKSUpQaN7dTm9lWR8QbZ6Zad7+SUAAOLE2fZcz5b0U9/H/y7pL37PzZR0rnDdW1KJ3/1SSZMDrCvg1xpj5kqaK0kFBQUB/vUAol1acqIm+MJzg33llaf3bX+884j+uGSnHv/Q7WzL75LWZHV7dO8spSWzug0ACK6zhWvTysct3T/X6xsEugcl4Ndaax+V9KjktoUE+PcDiEH5WWm6akxPXTWmpySpurZeG/b5rW7vOqJX1+6TJCUnGo3s2aVJK8A+XdNlDINuAADtd7ZwbVv5uKX7LSmV1Nfvfh9JewKsqyOvBQBJbnvI2D7ZGtsnW7f5HjtQUdVkdfvpZSV6ctEOSVKPjNQmq9tj+2SpU0ogTZUAAHDO9lNjnDHmmNwqcrrvY/nuB3Jp/jJJQ4wxAyTtlttmclOAdXXktQDQqpzMVF0xKl9XjMqXJNXW1WvDvgqtKDmqFb4V7jeL90uSEhOMhudn+gXururfvROr2wCAVoWsW4gkGWOulvSAXDu9J6y1PzLGzJMka+0jxph8SUWSukiql3Rc0khr7bGWXnuuz0e3EADBcPhEtVaWNLYBXFVSruNVtZKkrp2ST7cAHF/QVeP6ZisjldVtAIgnnrTi8wLhGkAo1NVbbS6rcGHbt7q9pey4JMkYaVheZpO92wN7dFZCAqvbABCrCNcAEGTlp2q0ssSNb/9411Gt3HVExyrd6naXtCSd57e6fV7fbGWlJ3tcMQAgWDo6/hwA0ExWerIuGpqji4bmSJLq6622HTyuj3e5wL1i11HNf3uzGtYvBudmNBnjPjg3Q4msbgNAzGHlGgBCpKKyRqtLy09vJVmx64iOnKyRJGWmJemmyQWae8FAdc9I9bhSAEBbsC0EACKAtVY7Dp3Uil1H9M6GMr2yZq9SkxJ1y7R+hGwAiCKEawCIQFvKjuuhdzbrhVV7CNkAEEUI1wAQwQjZABBdCNcAEAUaQvbzq/YoPTlRt0ztr7suGEDIBoAIQ7gGgCiypey4HvStZBOyASDyEK4BIAoRsgEgMhGuASCKEbIBILIQrgEgBmwpq9CD72whZAOAxwjXABBDWgrZcy8cqG6dU7wuDQDiAuEaAGJQ85B967T+uusCQjYAhBrhGgBi2JayCi14e4teXE3IBoBwIFwDQBwgZANAeBCuASCOELIBILQI1wAQhwjZABAahGsAiGOEbAAILsI1AKBJyO7kC9lfIGQDQJsRrgEAp23eX6EF72zRS4RsAGgXwjUA4AyEbABoH8I1AKBVhGwAaBvCNQDgnFoK2XddMFBdCdkA0AThGgAQsOYh+7bp/fWFGYRsAGhAuAYAtBkhGwBaRrgGALTbpv0VWvD2Zr28Zi8hGwBEuAYABAEhGwAcwjUAIGgI2QDiHeEaABB0hGwA8YpwDQAIGf+Q3TklSbdN6687ZwwgZAOIWYRrAEDIEbIBxAvCNQAgbDbuq9CCdzbrFb+Q/YULBii7EyEbQGwgXAMAwo6QDSBWEa4BAJ4hZAOINYRrAIDnCNkAYgXhGgAQMZqH7NunuwsfCdkAogXhGgAQcRpC9sur9yojlZANIHoQrgEAEWvjvsYWfoRsANGAcA0AiHiEbADRgnANAIgahGwAkY5wDQCIOhv2HdODb28hZAOIOIRrAEDU8g/Zmb6QfQchG4CHCNcAgKhHyAYQKQjXAICYsWHfMS14e7NeWbPvdMi+c8ZAZXVK9ro0AHGCcA0AiDmEbABeIVwDAGIWIRtAuBGuAQAxj5ANIFwI1wCAuHFGyJ4xQHdOH0DIBhA0hGsAQNwhZAMIFcI1ACBurd/rQvarawnZAIKDcA0AiHuEbADBQrgGAMCHkA2gowjXAAA0Q8gG0F6EawAAWtEkZKcl6Y7pA3THjAHKSidkA2gZ4RoAgHMgZAMIFOEaAIAAFe9xIfu1dYRsAC0jXAMA0EaEbACtIVwDANBOhGwAzRGuAQDoIEI2gAaEawAAgqR5yL5zxgDdPp2QDcQTwjUAAEFGyAbiF+EaAIAQWbenXAve3qzX1+0nZANxgnANAECIEbKB+EG4BgAgTAjZQOwjXAMAEGaEbCB2Ea4BAPCIf8jukpakO2cM1G3T+xOygShGuAYAwGNrd7uQ/UYxIRuIdoRrAAAiREsh+/YZ/dUljZANRAvCNQAAEYaQDUQvwjUAABGKkA1EH8I1AAARjpANRA/CNQAAUaJ5yP7CBe7CR0I2EDkI1wAARJm1u8s1/+3NepOQDUQcwjUAAFGKkA1EHsI1AABRjpANRA7CNQAAMYKQDXiPcA0AQIwhZAPeIVwDABCj/EN2VnqyvjBjgG4lZAMhRbgGACDGrd1drgfe2qy31hOygVAjXAMAECdaCtm3Te+vTEI2EDRnC9cJIf7EM40xG40xW4wx97fwvDHGLPA9v9oYM8HvuR3GmDXGmJXGGBIzAAABGN07S4/dWqiXvjRDE/t30/++uUkzfvIPPfj2ZlVU1nhdHhDzQrZybYxJlLRJ0uWSSiUtkzTHWlvsd8zVkr4k6WpJkyXNt9ZO9j23Q1KhtfZgoJ+TlWsAAJpaU+r2ZLOSDQSPVyvXkyRtsdZus9ZWS1ooaVazY2ZJ+r11lkjKNsb0DGFNAADElTF93Er2i/exkg2EQyjDdW9JJX73S32PBXqMlfSGMWa5MWZua5/EGDPXGFNkjCk6cOBAEMoGACD2tBayH3qHkA0EUyjDtWnhseZ7UM52zHRr7QRJV0m61xhzYUufxFr7qLW20FpbmJOT0/5qAQCIA01Ddlf9/A1CNhBMoQzXpZL6+t3vI2lPoMdYaxv+LJP0N7ltJgAAIAhcyJ5IyAaCLJThepmkIcaYAcaYFEmzJb3Q7JgXJN3i6xoyRVK5tXavMaazMSZTkowxnSVdIWltCGsFACAuEbKB4EoK1V9sra01xtwn6XVJiZKesNauM8bM8z3/iKRX5DqFbJF0UtLtvpfnSfqbMaahxj9ba18LVa0AAMS7hpDtuots0s/f2KTffLBdd10wQLdOo7sIECiGyAAAgDM0hOy31pcpKz2ZkA34YUIjAABoF/+Qnd0pWXddMFC3TO1HyEZcI1wDAIAOWV16VPPf2qy3NxCyAcI1AAAICkI2QLgGAABBRshGPCNcAwCAkCBkIx4RrgEAQEi1FLJvndZfGakh6/oLeIZwDQAAwoKQjXhAuAYAAGG1quSo5r+9We8QshGDCNcAAMAThGzEIsI1AADwFCEbsYRwDQAAIgIhG7GAcA0AACIKIRvRjHANAAAikn/I7topWXddOFC3TCVkI7IRrgEAQERbWXJU89/apH9sPEDIRsQjXAMAgKhAyEY0IFwDAICoQshGJCNcAwCAqETIRiQiXAMAgKhGyEYkIVwDAICY0Dxkz71wkG6Z2k+dCdkII8I1AACIKSt2HdH8tzfrXUI2PEC4BgAAMYmQDS8QrgEAQEwjZCOcCNcAACAuELIRDoRrAAAQVwjZCCXCNQAAiEv+Ibtb5xTNvXCgbp5CyEbHEK4BAEBc+3jXEc1/a7Pe20TIRscRrgEAAETIRnAQrgEAAPwQstERhGsAAIAWELLRHoRrAACAsyBkoy0I1wAAAAFoHrLvvnCgbp7aT51SCNloRLgGAABog+U7XQu/9wnZaAHhGgAAoB0I2WgJ4RoAAKAD/EN294Y92YTsuEW4BgAACAJCNiTCNQAAQFARsuMb4RoAACAEmofsuy8aqM9PIWTHOsI1AABACC3feVgPvLVZH2w+SMiOA4RrAACAMCBkxwfCNQAAQBgRsmMb4RoAAMADhOzYRLgGAADwECE7thCuAQAAIoB/yO6RkaK7Lxykz00pIGRHGcI1AABABCnacVjz3yZkRyvCNQAAQAQiZEcnwjUAAEAEI2RHF8I1AABAFCBkRwfCNQAAQBRpKWR/fko/packel0aRLgGAACISs1D9ryLBulzkwnZXiNcAwAARLFlOw5r/lub9eEWQnYkIFwDAADEAEJ2ZCBcAwAAxBBCtrcI1wAAADGIkO0NwjUAAEAMaxqyUzXvooGE7BAiXAMAAMQBQnZ4EK4BAADiyNLthzX/7U3655ZDhOwQIFwDAADEIUJ2aBCuAQAA4hghO7gI1wAAACBkBwnhGgAAAKf5h+yczFRfC78CpSUTsgNBuAYAAMAZCNntQ7gGAABAqz7adkjz396sRVsJ2YEgXAMAAOCcCNmBIVwDAAAgYITssyNcAwAAoM0I2S0jXAMAAKDdmofsL140SDfFccgmXAMAAKDDCNkO4RoAAABBs2TbIc1/a7MWb4vPkE24BgAAQNDFa8gmXAMAACBk4i1kE64BAAAQcvESsgnXAAAACBv/kJ2bmaovXjxIcybFTsgmXAMAACDsYjVkE64BAADgmcVbD2n+25u0ZNvhmAjZhGsAAAB4LlZCNuEaAAAAESPaQzbhGgAAABEnWkM24RoAAAARq3nIvufiQZodwSGbcA0AAICIt3jrIT3w1iZ9tD2yQzbhGgAAAFEj0kP22cJ1Qog/8UxjzEZjzBZjzP0tPG+MMQt8z682xkwI9LUAAACITVMHddfTd0/VU3dN0YAenfXdF4t10c/+oSf/uV2VNXVel3dWIVu5NsYkStok6XJJpZKWSZpjrS32O+ZqSV+SdLWkyZLmW2snB/LalrByDQAAEHsWbz2k/3trk5ZuP6y8Lm6supcr2V6tXE+StMVau81aWy1poaRZzY6ZJen31lkiKdsY0zPA1wIAACAOTB3UXc/4VrL7dW9cyX5h1R6vSztDUgj/7t6SSvzul8qtTp/rmN4BvhYAAABxZOqg7po6aOrplexIFMpwbVp4rPkelNaOCeS17i8wZq6kuZJUUFDQlvoAAAAQhRpCdiQ25gjltpBSSX397veR1HztvrVjAnmtJMla+6i1ttBaW5iTk9PhogEAABAdjGlpPdZboQzXyyQNMcYMMMakSJot6YVmx7wg6RZf15ApksqttXsDfC0AAAAQUUK2LcRaW2uMuU/S65ISJT1hrV1njJnne/4RSa/IdQrZIumkpNvP9tpQ1QoAAAAEA0NkAAAAgDbwbIgMAAAAEE8I1wAAAECQEK4BAACAICFcAwAAAEFCuAYAAACChHANAAAABAnhGgAAAAgSwjUAAAAQJIRrAAAAIEgI1wAAAECQEK4BAACAICFcAwAAAEFCuAYAAACChHANAAAABImx1npdQ9AYYw5I2ul1HVAPSQe9LgIRiXMDreHcQGs4N3A2Xp0f/ay1OS09EVPhGpHBGFNkrS30ug5EHs4NtIZzA63h3MDZROL5wbYQAAAAIEgI1wAAAECQEK4RCo96XQAiFucGWsO5gdZwbuBsIu78YM81AAAAECSsXAMAAABBQrhGuxhjZhpjNhpjthhj7m/h+c8ZY1b7bouMMeO8qBPeONf54XfcRGNMnTHm+nDWB+8Ecm4YYy42xqw0xqwzxrwX7hrhjQB+rmQZY140xqzynRu3e1Enws8Y84QxpswYs7aV540xZoHv3FltjJkQ7hqb1MO2ELSVMSZR0iZJl0sqlbRM0hxrbbHfMdMkrbfWHjHGXCXpu9bayZ4UjLAK5PzwO+5NSZWSnrDWPhvuWhFeAX7vyJa0SNJMa+0uY0yutbbMk4IRNgGeG9+WlGWt/ZYxJkfSRkn51tpqL2pG+BhjLpR0XNLvrbWjW3j+aklfknS1pMmS5nuZOVi5RntMkrTFWrvN901toaRZ/gdYaxdZa4/47i6R1CfMNcI75zw/fL4k6a+SCE7xI5Bz4yZJz1lrd0kSwTpuBHJuWEmZxhgjKUPSYUm14S0TXrDWvi/3792aWXLB21prl0jKNsb0DE91ZyJcoz16Syrxu1/qe6w1d0p6NaQVIZKc8/wwxvSW9C+SHgljXfBeIN87hkrqaox51xiz3BhzS9iqg5cCOTcekjRC0h5JayR9xVpbH57yEOHamktCKsmrT4yoZlp4rMX9RcaYS+TC9YyQVoRIEsj58YCkb1lr69wiFOJEIOdGkqTzJV0qKV3SYmPMEmvtplAXB08Fcm5cKWmlpE9IGiTpTWPMB9baY6EuDhEv4FwSDoRrtEeppL5+9/vIrSQ0YYwZK+kxSVdZaw+FqTZ4L5Dzo1DSQl+w7iHpamNMrbX27+EpER4J5NwolXTQWntC0gljzPuSxsntx0XsCuTcuF3Sj627WGyLMWa7pOGSloanRESwgHJJuLAtBO2xTNIQY8wAY0yKpNmSXvA/wBhTIOk5STez4hR3znl+WGsHWGv7W2v7S3pW0j0E67hwznND0vOSLjDGJBljOsldnLQ+zHUi/AI5N3bJvaMhY0yepGGStoW1SkSqFyTd4usaMkVSubV2r1fFsHKNNrPW1hpj7pP0uqREuU4P64wx83zPPyLpvyR1l/Qr3+pkrbW20KuaET4Bnh+IQ4GcG9ba9caY1ySt7THAZAAAAeNJREFUllQv6TFrbYvttxA7/n97d6giVRyFAfw7QXyBfQODTQSTbFkRljUI4hv4DIIGBYN2H0CQDcKaRUTFZDEYFDTJxi2KCIIoYjiGqzIuEwzXGebO79fu/074wnD5mDnc84/PjZtJdqvqTYYxgKvd/XFpoVmYqtpLspVko6oOktxIciT58914lOFNIftJvmb4l2NpvIoPAABGYiwEAABGolwDAMBIlGsAABiJcg0AACNRrgEAYCTKNcDEVdXdqvpQVW8PnZ+uqjtVtVVVD2fOb1XVk6o6uvi0AKtNuQaYvt0kO3POd5I8nj2oqmtJNpNc6O7v/z8awLQo1wAT193Pk3yac+tskme/L6rqcoZFDOe7+9uC4gFMig2NAGuoqjaS/Ojuz7+2qG5mWCd9qru/LDUcwArzyzXAetpO8nTmej/DSunt5cQBmAblGmA9ncvf89bvM4yE3K6qM8uJBLD6lGuANVPDHMiJJK9nz7v7XZKLSe5V1cllZANYdco1wMRV1V6SF0mOV9VBkitJXnV3H/5sd79McinJg6o6ttikAKuv5jxbAZiwqrqeZL+77y87C8DUKNcAADASYyEAADAS5RoAAEaiXAMAwEiUawAAGIlyDQAAI1GuAQBgJMo1AACM5Ccw/t1A2Z76RQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot knn_cv_score vs k_value to find best k_value\n",
    "plt.figure(figsize=(12, 8))\n",
    "plt.plot([1/k for k in k_value], [1-x for x in training_accuracy])\n",
    "plt.plot([1/k for k in k_value], [1-x for x in testing_accuracy])\n",
    "plt.xlabel('1/K')\n",
    "plt.ylabel('Error rate')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "SkiZBkBbPYAV",
    "outputId": "13777844-51bd-4366-d260-ecb89efc7b83"
   },
   "source": [
    "From the two plots above, the best value of k to use is 3 for our KNN Model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=3)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# creat KNN model\n",
    "knn_model = KNeighborsClassifier(n_neighbors=3)\n",
    "\n",
    "# fit model with training data\n",
    "knn_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN Train Set Accuracy: 0.8959904761904762\n",
      "KNN Test Set Accuracy: 0.8089444444444445\n"
     ]
    }
   ],
   "source": [
    "# display training and test accuracy\n",
    "print(f'KNN Train Set Accuracy: {knn_model.score(X_train, y_train)}')\n",
    "print(f'KNN Test Set Accuracy: {knn_model.score(X_test, y_test)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Question 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our best model is the decision tree. Use the parameters from above to create decision tree model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(max_depth=19)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# create decision tree model\n",
    "dt_model = DecisionTreeClassifier(max_depth=19,\n",
    "min_samples_leaf=1)\n",
    "\n",
    "# fit decision tree model\n",
    "dt_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DT Model Training Accuracy Score: 0.9567476190476191\n"
     ]
    }
   ],
   "source": [
    "# train accuracy score\n",
    "dt_train_accuracy = dt_model.score(X_train, y_train)\n",
    "\n",
    "# display training accuracy score\n",
    "print(f'DT Model Training Accuracy Score: {dt_train_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DT Model Test Accuracy Score: 0.9023888888888889\n"
     ]
    }
   ],
   "source": [
    "# testing accuracy score\n",
    "dt_test_accuracy = dt_model.score(X_test, y_test)\n",
    "\n",
    "# display test accuracy score\n",
    "print(f'DT Model Test Accuracy Score: {dt_test_accuracy}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Question 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plotting a figure that demonstrates which features are most important in constructing more accurate predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "# retrieve our feature importance from the model\n",
    "feature_importance = pd.Series(dt_model.feature_importances_, index=X_train.columns).sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAykAAAHSCAYAAADooISZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de7RdZX3v//eHQMM9qICNVA1gDt7QAAFFkdZKrTVVREE4ReXirwx/4rUDT/F4o/rzNKe2RyzWan5WQQ+11ALKr7SIUrmICNmBQEDBG6EttXqkGgW8QPj+/lgzZbGzd/ba2XtlPcl+v8bYY6/1zOeZ87ueTBj55JlzzVQVkiRJktSK7UZdgCRJkiT1M6RIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkpmw/6gLUnj333LMWLVo06jIkSZK0jVu1atUPq2qv8e2GFG1k0aJFjI2NjboMSZIkbeOS3DVRu5d7SZIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqiiFFkiRJUlO2H3UBas+au9ex6MxLJ92+dvmyLViNJEmS5hpXUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqiiFlBpLskeT1W+hYpyZZk+SWJLcmObprvzLJ0r5+i7rtuyS5J8mCcfv5XJJXbomaJUmSpM1hSJmZPYChh5Qkvwa8Aziiqp4BPBu4ZVNjquo+4HLgZX37WQAcAfz98KqVJEmSZsaQMjPLgf2TrE7y2Q2rGwBJzk/y0iQnJ/l8ksuS3JHkPX19XpXkhm78x5LMm+Q4ewM/Be4FqKp7q+rOAer7DHBC3/tjgMuq6v7pflBJkiRpSzGkzMyZwHeqagnwYeAU+M8Vi+cA/9D1Oww4EVgCHJdkaZKnAMcDz+3Gr+/6TORm4PvAnUk+meQlA9Z3GXBIksd070+gF1w2kuS0JGNJxtbfv27A3UuSJEmzz5AyS6rqKuBJSfYG/itwYVU92G3+YlXdU1U/Ay6id8nVC4BDgJVJVnfv95tk3+uBFwHHAt8EPpjkrA2bJxrSjfslcAlwbJI96YWkyyc5xoqqWlpVS+ftvGCiLpIkSdIWsf2oC9jGfJreasgJwKl97eODRAEBzquqtw+y46oq4AbghiRfBD4JnAXcAzyqr+ujgR/2vf8M8M7ueJ+vqgcG/TCSJEnSKLiSMjM/BXbre38u8BaAqrqtr/23kjw6yU70bmS/FriC3grH3gDd9idOdJAkj0tycF/TEuCu7vWVwKuSpHt/EvDlvr5fBhYDpzPJpV6SJElSS1xJmYGquifJtUluBf6xqt6W5BvA58Z1/Qq9VZYnAX9dVWMASd4JXJ5kO+ABekHiLja2A/CnSR4H/Bz4P8Drum0rgCcDNycpYAz4z9WZqnooyYXAccDVs/G5JUmSpGEypMxQVf3ehtdJdqa3ajF+xeIHVfWGCcZeAFwwwDHuAn5zkm2/BDba97g+bwbePNVxJEmSpBZ4udcsSXIUcDtwTlX59ViSJEnSZnIlZZZU1ZeAJ0zQfi69e1UGkuR6YP645ldX1ZqZ1CdJkiRtLQwpjamqZ426BkmSJGmUvNxLkiRJUlMMKZIkSZKaYkiRJEmS1BRDiiRJkqSmGFIkSZIkNcWQIkmSJKkphhRJkiRJTfE5KdrIgfssYGz5slGXIUmSpDnKlRRJkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xeekaCNr7l7HojMvnXT7Wp+hIkmSpCFyJUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkpmwzISXJXkmuT3JTkudNc+ySJC8eVm2bOO5/H8I+FyX5vdneryRJkrSlbDMhBXgBcHtVHVRV10xz7BJgWiElPVPO3xT9Zj2kAIsAQ4okSZK2WkMLKd2/6N+e5ONJbk1yfpKjklyb5FtJDut+vtqtfnw1yQHd2JOTXJTksq7vn/Tt996+18cmOTfJEuBPgBcnWZ1kpyR/mWQsyW1J/qhvzKHdsW5OckOSBcB7geO7sccnOSvJGX1jbu0+z6Ik30jyEeBG4PFJ3pZkZZJbNhxnon4TzM9yYKfumOd3bX/QHevWJG/p2t6X5M19496f5E1d+PlA13dNkuO7LsuB53X7fWtXyzVJbux+njPDP1pJkiRpqLYf8v6fBBwHnAaspPcv/EcAL6W3ivAa4MiqejDJUcD/AF7RjV0CHAT8ArgjyTlV9S8THaSqVid5N7C0qt4AkOQdVfUfSeYBVyR5BnA7cAFwfFWtTLI7cD8wfuxZm/hMBwCnVNXrk7wQWAwcBgS4JMmRwD/395uk5jOTvKGqlnTHPAQ4BXhWt6/rk1wF/BVwEfChbkXmhO54L+/m6JnAnsDKJFcDZwJnVNXvdvvdGfitqvp5ksXAZ4Cl4+tJchq9Pyfm7b7XJj6+JEmSNFzDDil3VtUagCS3AVdUVSVZQ++ypAXAed1fngvYoW/sFVW1rhv7deCJwIQhZRKv7P7ivT2wEHhqd4zvVdVKgKr6Sbf/6Xymu6rqa93rF3Y/N3Xvd6UXWv55XL9BHAFcXFX3dTVdBDyvqv48yT1JDgIeC9xUVfckOQL4TFWtB77fBZpDgZ+M2+8OwIe71ab1wH+Z6OBVtQJYATB/4eKaRt2SJEnSrBp2SPlF3+uH+t4/1B37fcCXq+qYJIuAKycZu56Ha+3/C/SOEx00yb7AGcChVfWjJOd2fTNu/GQe5JGXwvUf577+QwF/XFUfG3f8ReP6DWJTSenjwMnArwKfGKB/v7cC36e34rId8PNp1iVJkiRtUaO+cX4BcHf3+uQBx3w/yVO6S5+OmaTP7vRCwrokjwV+p2u/HXhckkMBkuyWZHvgp8BufePXAgd3fQ4G9p3kOF8ATk2ya9d3nyR7D/g5AB5IsmH16GrgZUl2TrJL99k2fAHAxcCL6K2UfKGv//FJ5iXZCzgSuGGCz7KA3urRQ8CrgXnTqE+SJEna4oa9kjKVP6F3udcfAP804Jgzgb+nd+nXrfQusXqEqro5yU3AbcB3gWu79l92N5ifk2Qn4GfAUcCXgTOTrAb+GLgQeE33fiXwzYkKqarLkzwFuK67ZOxe4FX0Vn4GsQK4JcmNVXVit+JzQ7ft41V1U1/dXwZ+3F3eBb3gcjhwM73Vof9WVf+e5B7gwSQ3A+cCHwEuTHJc9zmnu8IjSZIkbVGp8vaD1nWrRjcCx1XVt4Z9vPkLF9fCk86edPva5cuGXYIkSZLmgCSrqmqjL3Ua9eVemkKSpwLfpvdFAkMPKJIkSdKojfpyrzkhyfXA/HHNr97wzWebUlVfB/YbSmGSJElSgwwpW0BVPWvUNUiSJElbCy/3kiRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqiiFFkiRJUlN8Too2cuA+CxhbvmzUZUiSJGmOciVFkiRJUlMMKZIkSZKaYkiRJEmS1BRDiiRJkqSmGFIkSZIkNcWQIkmSJKkphhRJkiRJTfE5KdrImrvXsejMSwfuv9ZnqkiSJGkWuZIiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqiiFFkiRJUlMMKZIkSZKaYkiZQJI9krx+1HVMJcm9o65BkiRJmm2GlIntATQfUiRJkqRtkSFlYsuB/ZOsTvLZJEdv2JDk/CQvTXJyks8nuSzJHUne09fnVUlu6MZ/LMm8iQ6SZF6Sc5PcmmRNkrd27ft3+12V5JokT+7a901yXZKVSd7Xt59dk1yR5MZuP0d37YuSfCPJ/5vktiSXJ9lpSHMmSZIkzQpDysTOBL5TVUuADwOnACRZADwH+Ieu32HAicAS4LgkS5M8BTgeeG43fn3XZyJLgH2q6ulVdSDwya59BfDGqjoEOAP4SNf+IeAvq+pQ4N/79vNz4JiqOhh4PvBnSdJtWwz8RVU9Dfgx8IqJCklyWpKxJGPr7183wBRJkiRJw7H9qAtoXVVdleQvkuwNvBy4sKoe7DLAF6vqHoAkFwFHAA8ChwAruz47AT+YZPffBfZLcg5wKXB5kl3pBaHPPpwzmN/9fi4Ph4xPA/+zex3gfyQ5EngI2Ad4bLftzqpa3b1eBSya5HOuoBeOmL9wcU0xLZIkSdLQGFIG82l6qyEnAKf2tY//y3zRCwznVdXbp9ppVf0oyTOB3wZOB14JvAX4cbcKM+GwCdpOBPYCDqmqB5KsBXbstv2ir996eqFJkiRJapaXe03sp8Bufe/PpRceqKrb+tp/K8mju/s8XgZcC1wBHNutvNBtf+JEB0myJ7BdVV0IvAs4uKp+AtyZ5LiuT7ogQ7f/E7rX/ZeQLQB+0AWU5wMTHk+SJEnaGhhSJtBdwnVtd0P7B6rq+8A3ePiekQ2+Qm+VZTW9y8DGqurrwDvpXbp1C/BFYOEkh9oHuDLJanpBaMPqy4nAa5PcDNwGbLhx/83A6UlW0gsmG5wPLE0y1o29fTM/uiRJkjRyqfL2g6kk2RlYQ2+lY13XdjKwtKreMMrahmH+wsW18KSzB+6/dvmyIVYjSZKkbVWSVVW1dHy7KylTSHIUvZWJczYEFEmSJEnD443zU6iqLwFPmKD9XHqXaA0kyfU8/C1dG7y6qtbMpD5JkiRpW2NI2UKq6lmjrkGSJEnaGni5lyRJkqSmGFIkSZIkNcWQIkmSJKkphhRJkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkpric1K0kQP3WcDY8mWjLkOSJElzlCspkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaorPSdFG1ty9jkVnXjqtMWt9rookSZJmiSspkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkphhSJEmSJDXFkNJJskeS14+4hvcmOWoT209O8uEtWZMkSZK0pRlSHrYHMNKQUlXvrqovjbIGSZIkadQMKQ9bDuyfZHWSzyY5esOGJOcneWm3kvH5JJcluSPJe/r6vCrJDd34jyWZN9mBktyb5M+S3JjkiiR7de3nJjm2e31okq8mubnb727j9rEsyXVJ9kyyNsmeXfvSJFd2r89K8ukk/5TkW0l+fzYnTJIkSRoGQ8rDzgS+U1VLgA8DpwAkWQA8B/iHrt9hwInAEuC4LhQ8BTgeeG43fn3XZzK7ADdW1cHAVcB7+jcm+RXgAuDNVfVM4CjgZ33bj+nqfXFV/XCKz/UMYBlwOPDuJI+bqFOS05KMJRlbf/+6KXYpSZIkDc/2oy6gRVV1VZK/SLI38HLgwqp6MAnAF6vqHoAkFwFHAA8ChwAruz47AT/YxCEeohdCAP43cNG47QcA36uqlV09P+mOB/B8YCnwwg3tU/h8Vf0M+FmSL9MLWZ+b4DOvAFYAzF+4uAbYryRJkjQUhpTJfZreasgJwKl97eP/Al9AgPOq6u2beazx+8wEbRt8F9gP+C/AWNf2IA+viu04xb4NIJIkSWqal3s97KdA/30f5wJvAaiq2/rafyvJo5PsBLwMuBa4Aji2W3mh2/7ETRxrO+DY7vXvAV8Zt/124HFJDu32t1uSDYHyLnqrO59K8rSubS29lRyAV4zb19FJdkzyGOA3gJWbqEuSJEkaOVdSOlV1T5Jrk9wK/GNVvS3JN9j40qiv0FtleRLw11U1BpDkncDlSbYDHgBOpxcoJnIf8LQkq4B19O5n6a/ll0mOB87pwtDP6N2XsmH7HUlOBD6b5CXAHwF/leS/A9ePO9YNwKXAE4D3VdW/TWNaJEmSpC0uVV79M5EkOwNrgIOral3XdjKwtKreMMN931tVu868yimPcxZwb1X96XTGzV+4uBaedPa0jrV2+bJp9ZckSZKSrKqqpePbvdxrAt0DFW8HztkQUCRJkiRtGV7uNYHugYpPmKD9XHr3qgwkyfXA/HHNr94SqygAVXXWljiOJEmSNJsMKUNUVc8adQ2SJEnS1sbLvSRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUn5OijRy4zwLGli8bdRmSJEmao1xJkSRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqiiFFkiRJUlN8Too2subudSw689JRlzFSa31OjCRJ0si4kiJJkiSpKYYUSZIkSU0xpEiSJElqiiFFkiRJUlMMKZIkSZKaYkiRJEmS1BRDiiRJkqSmGFIkSZIkNcWQIkmSJKkphhRJkiRJTTGkSJIkSWqKIUWSJElSUwwpA0rypiTfSHL+Zoxdm2TPAfueleSM7vV7kxw13eNtYt+/keTvZ2t/kiRJ0jBsP+oCtiKvB36nqu7cUgesqndvqWNJkiRJrXAlZQBJPgrsB1ySZN2GlY5u261JFnWvX5XkhiSrk3wsybwB9/+OJHck+RJwQF/7uUmO7V4fkuSqJKuSfCHJwq79TUm+nuSWJH/Tte2S5BNJVia5KcnRszQVkiRJ0tAZUgZQVa8D/g14PvDBifokeQpwPPDcqloCrAdOnGrfSQ4BTgAOAl4OHDpBnx2Ac4Bjq+oQ4BPA+7vNZwIHVdUzgNd1be8A/qmqDu1q/kCSXaao47QkY0nG1t+/bqqyJUmSpKEZ6HKvJPsD/1pVv0jyG8AzgE9V1Y+HWdxW5gXAIcDKJAA7AT8YYNzzgIur6n6AJJdM0OcA4OnAF7t9zwO+1227BTg/yeeAz3VtLwRe2rfisyPwhE0VUVUrgBUA8xcurgHqliRJkoZi0HtSLgSWJnkS8FfAJcBfAy8eVmENe5BHrkDt2P0OcF5VvX0z9jlVKAhwW1UdPsG2ZcCRwEuBdyV5Wtf/FVV1xyN2kjx2M2qTJEmStqhBL/d6qKoeBI4Bzq6qtwILh1dW09YCBwMkORjYt2u/Ajg2yd7dtkcneeIA+7saOCbJTkl2A14yQZ87gL2SHN7te4ckT0uyHfD4qvoy8N+APYBdgS8Ab0y37JLkoM37qJIkSdKWN+hKygNJ/itwEg//JXqH4ZTUvAuB1yRZDawEvglQVV9P8k7g8i48PACcDty1qZ1V1Y1JLgBWd32vmaDPL7sb6P88yQJ6f25nd8f+311bgA9W1Y+TvK/bfksXVNYCvzvzjy5JkiQNX6qmvv0gyVPp3ZR9XVV9Jsm+wPFVtXzYBWrLm79wcS086exRlzFSa5cvG3UJkiRJ27wkq6pq6fj2gVZSulWCP6S7+bp7VogBRZIkSdKsG+ielCQvoXc50mXd+yWTfAuVJpHkMd3zU8b/PGbUtUmSJEktGfSelLOAw4ArAapqdXfJlwZUVfcAS0ZdhyRJktS6Qb/d68GqGv+EP5+lIUmSJGnWDbqScmuS3wPmJVkMvAn46vDKkiRJkjRXDbqS8kbgacAv6D3EcR3wlmEVJUmSJGnumnIlJck84JKqOgp4x/BLkiRJkjSXTbmSUlXrgfu7BwZKkiRJ0lANek/Kz4E1Sb4I3LehsareNJSqJEmSJM1Zg4aUS7sfSZIkSRqqVPlNwnqkpUuX1tjY2KjLkCRJ0jYuyaqqWjq+faCVlCR3MsFzUapqv1moTZIkSZL+06CXe/Wnmx2B44BHz345kiRJkua6gZ6TUlX39P3cXVVnA7855NokSZIkzUGDXu51cN/b7eitrOw2lIokSZIkzWmDXu71Z32vHwTuBF45++VIkiRJmusGDSmvrarv9jck2XcI9UiSJEma4wa6JwX4uwHbJEmSJGlGNrmSkuTJwNOABUle3rdpd3rf8qVt0Jq717HoTJ/dOYi1y5eNugRJkqRtzlSXex0A/C6wB/CSvvafAr8/rKIkSZIkzV2bDClV9Xng80kOr6rrtlBNkiRJkuawQW+cvynJ6fQu/frPy7yq6tShVCVJkiRpzhr0xvlPA78K/DZwFfBr9C75kiRJkqRZNWhIeVJVvQu4r6rOA5YBBw6vLEmSJElz1aAh5YHu94+TPB1YACwaSkWSJEmS5rRB70lZkeRRwLuAS4BdgXcPrSpJkiRJc9ZAIaWqPt69vArYb3jlSJIkSZrrBrrcK8ljk/xVkn/s3j81yWuHW5okSZKkuWjQe1LOBb4APK57/03gLcMoSJIkSdLcNmhI2bOq/hZ4CKCqHgTWD60qSZIkSXPWoCHlviSPAQogybOBdUOrSpIkSdKcNWhI+QN63+q1f5JrgU8BbxxaVduwJOcmObZ7/fEkT53m+HuHU5kkSZLUhk1+u1eSJ1TVP1fVjUl+HTgACHBHVT2wqbGaWlX9X8Pcf5IAqaqHhnkcSZIkaTZNtZLyub7XF1TVbVV1qwFlY0lek+SWJDcnuTjJnUl26LbtnmTthvd9Y65MsrR7fW+S93fjv5bksV37vkmuS7IyyfvGjX9b135Lkj/q2hYl+UaSjwA3Ao/vVm9uTbImyVu3xHxIkiRJm2uqkJK+1z4fZRJJnga8A/jNqnom8FrgSmBZ1+UE4MIpwt0uwNe68VcDv9+1fwj4y6o6FPj3vmO+EFgMHAYsAQ5JcmS3+QDgU1V1ELAnsE9VPb2qDgQ+OdPPK0mSJA3TVCGlJnmtR/pN4O+q6ocAVfUfwMeBU7rtpzB1OPgl8Pfd61XAou71c4HPdK8/3df/hd3PTfRWTJ5ML7QA3FVVX+tefxfYL8k5SV4E/GSigyc5LclYkrH19/udCJIkSRqdqZ44/8wkP6G3orJT95rufVXV7kOtbusRxoW4qrq2u/Tq14F5VXXrFPt4oKo27GM9j/yzmSggBvjjqvrYIxqTRcB9fXX8KMkzgd8GTgdeCZw6fmdVtQJYATB/4WIDqSRJkkZmkyspVTWvqnavqt2qavvu9Yb3BpSHXQG8svuaZpI8umv/FL1VkJlcYnUtvcvFAE7sa/8CcGqSXbtj7pNk7/GDk+wJbFdVFwLvAg6eQS2SJEnS0A36FcTahKq6DXg/cFWSm4H/1W06H3gUD1+utTneDJyeZCWwoO+YlwN/DVyXZA3wd8BuE4zfB7gyyWrgXODtM6hFkiRJGro8fIWRZlv3PJSjq+rVo65lOuYvXFwLTzp71GVsFdYuXzZ1J0mSJE0oyaqqWjq+fap7UrSZkpwD/A7w4lHXIkmSJG1NDClDUlVvHHUNkiRJ0tbIe1IkSZIkNcWQIkmSJKkphhRJkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xSfOayMH7rOAseXLRl2GJEmS5ihXUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUn5Oijay5ex2Lzrx01GWos9Zn1kiSpDnGlRRJkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJImSVJzkpyxgTtr0vymlk8zpOTrE5yU5L9Z6NGSZIkqSXbj7qAbVmS7avqo7O825cBn6+q98zyfiVJkqQmGFJmIMk7gNcA/wL8H2BVkiuBrwLPBS5JshtwL3ApcF5VHdaNXQRcUlXPSHII8L+AXYEfAidX1fcmON6LgbcA65McCXwDuKyqLklyMfCjqjo1yWuBfavqnRPVOJzZkCRJkmaHl3ttpi5YnAAcBLwcOLRv8x5V9etV9WcbGqrqG8CvJNmvazoe+NskOwDnAMdW1SHAJ4D3T3TMqvoH4KPAB6vq+cDVwPO6zfsAT+1eHwFcM0WN4z/PaUnGkoytv3/doNMgSZIkzTpDyuZ7HnBxVd1fVT8BLunbdsEkY/4WeGX3+viu3wHA04EvJlkNvBP4tQFruAZ4XpKnAl8Hvp9kIXA4vdWcTdX4CFW1oqqWVtXSeTsvGPDwkiRJ0uzzcq+ZqUna75uk/QLgs0kuAqqqvpXkQOC2qjp82gevujvJo4AX0VtVeTS9EHRvVf00yaZqlCRJkprkSsrmuxo4JslO3X0nL5lqQFV9B1gPvIuHV1vuAPZKcjhAkh2SPG0adVxH7z6Vq+mtrJzR/d6sGiVJkqRRcyVlM1XVjUkuAFYDd/FwMJjKBcAHgH27/fwyybHAnydZQO/P5GzgtgH3dw3wwqr6dpK76K2mXDPDGiVJkqSRSZVXA+mR5i9cXAtPOnvUZaizdvmyUZcgSZI0FElWVdXS8e1e7iVJkiSpKV7u1agkf0HvWSv9PlRVnxxFPZIkSdKWYkhpVFWdPuoaJEmSpFHwci9JkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xeekaCMH7rOAseXLRl2GJEmS5ihXUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUn5Oijay5ex2Lzrx01GVoM631GTeSJGkr50qKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU1pNqQkOSvJGUPa93uTHDVFn5cleep0xkyzhpOTfHi29jduv4/re782yRqXD4IAAA8HSURBVJ6zfRxJkiRpWLYfdQFbWpJ5VfXuAbq+DPh74OsAA45pwcnArcC/jbgOSZIkabM0tZKS5B1J7kjyJeCArm3/JJclWZXkmiRP7tqPS3JrkpuTXN21zUvyp0nWJLklyRu79rVJ3p3kK8BxSc5Ncmzftv+Z5Ibu50lJngO8FPhAktVdDf1jXpDkpu44n0gyv29ff5Tkxm7bkwf83HsluTDJyu7nuV37Wd3+r0zy3SRv6hvzriS3J/liks8kOaOrbylwflf3Tl33N063JkmSJGlUmgkpSQ4BTgAOAl4OHNptWgG8saoOAc4APtK1vxv47ap6Jr1AAXAasC9wUFU9Azi/7xA/r6ojqupvJjj8T6rqMODDwNlV9VXgEuBtVbWkqr7TV+eOwLnA8VV1IL3VqP+7b18/rKqDgb/s6h3Eh4APVtWhwCuAj/dtezLw28BhwHuS7JBkaddvw1wtBaiqvwPGgBO7un82aE1JTksylmRs/f3rBixbkiRJmn3NhBTgecDFVXV/Vf2EXkjYEXgO8Nkkq4GPAQu7/tcC5yb5fWBe13YU8NGqehCgqv6jb/8XbOLYn+n7ffgUdR4A3FlV3+zenwcc2bf9ou73KmDRFPva4Cjgw91nvATYPclu3bZLq+oXVfVD4AfAY4EjgM9X1c+q6qfA/zfF/qesqapWVNXSqlo6b+cFA5YtSZIkzb7W7kmpce+3A35cVUs26lj1uiTPApYBq5MsATLBPja4b8DjTjZ+g0yx/Rfd7/UMPr/bAYf3rXz0DpT0769/n1PVMBs1SZIkSSPR0krK1cAxSXbqVhFeAtwP3JnkOID0PLN7vX9VXd/d0P5D4PHA5cDrkmzf9Xn0gMc+vu/3dd3rnwK7TdD3dmBRkid1718NXDWNzzmRy4E3bHjTBa5N+QrwkiQ7JtmVXlDbYLK6JUmSpK1CMyGlqm6kd0nWauBC4Jpu04nAa5PcDNwGHN21f6C7EfxWegHnZnr3cvwzcEvX//cGPPz8JNcDbwbe2rX9DfC27gb5/fvq/DlwCr1L0NYADwEf3ZzP3OdNwNLuZv+vA6/bVOeqWknvsrCb6V3KNQZsuJHkXOCj426clyRJkrYaqZrq6qZtW5K1wNLuno+tRpJdq+reJDvTC2mndUFvxuYvXFwLTzp7NnalEVi7fNnUnSRJkhqQZFVVLR3f7v0JW68V3cMmdwTOm62AIkmSJI3anA8pVbVomPtPcgq9y8j6XVtVp89kv1U16KVskiRJ0lZlzoeUYauqTwKfHHUdkiRJ0taimRvnJUmSJAkMKZIkSZIaY0iRJEmS1BRDiiRJkqSmGFIkSZIkNcWQIkmSJKkphhRJkiRJTfE5KdrIgfssYGz5slGXIUmSpDnKlRRJkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xeekaCNr7l7HojMvHXUZ0jZjrc8dkiRpWlxJkSRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqiiFFkiRJUlMMKZIkSZKaYkiRJEmS1BRDiiRJkqSmGFIkSZIkNcWQIkmSJKkphpQhSnJWkjOGfIz/Psz9S5IkSVuaIWXrN+2QkmTeMAqRJEmSZoMhZQaS/EGSW7uft3Rt70hyR5IvAQf09V2S5GtJbklycZJHde1XJvlgkquTfCPJoUkuSvKtJP9P3/hXJbkhyeokH0syL8lyYKeu7fzJ+nXt9yZ5b5LrgcO34DRJkiRJ02JI2UxJDgFOAZ4FPBv4/a7tBOAg4OXAoX1DPgX8YVU9A1gDvKdv2y+r6kjgo8DngdOBpwMnJ3lMkqcAxwPPraolwHrgxKo6E/hZVS2pqhMn69cdYxfg1qp6VlV9ZdYnRJIkSZol24+6gK3YEcDFVXUfQJKLgGVd2/1d2yXd7wXAHlV1VTf2POCzffu6pPu9Britqr7Xjfsu8PjuWIcAK5MA7AT8YIKaXrCJfuuBCyf7MElOA04DmLf7XgNNgCRJkjQMhpTNl0naazP29Yvu90N9rze837471nlV9fYBapqs38+rav1kA6tqBbACYP7CxZvzGSRJkqRZ4eVem+9q4GVJdk6yC3AMcClwTJKdkuwGvASgqtYBP0ryvG7sq4GrJtrpJK4Ajk2yN0CSRyd5YrftgSQ7DNBPkiRJ2iq4krKZqurGJOcCN3RNH6+qVUkuAFYDdwHX9A05Cfhokp2B79K7n2XQY309yTuBy5NsBzxA776Vu+itftyS5MbuvpTJ+kmSJElbhVR5ZY8eaf7CxbXwpLNHXYa0zVi7fNmoS5AkqUlJVlXV0vHtXu4lSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkphhSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqyvajLkDtOXCfBYwtXzbqMiRJkjRHuZIiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkpvicFG1kzd3rWHTmpaMuQ5IkSUO2ttFn47mSIkmSJKkphhRJkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDClzUJJ5o65BkiRJmowhpXFJ3pfkzX3v35/kTUnelmRlkluS/FHf9s8lWZXktiSn9bXfm+S9Sa4HDt/CH0OSJEkamCGlfX8FnASQZDvgBOD7wGLgMGAJcEiSI7v+p1bVIcBS4E1JHtO17wLcWlXPqqqvbMkPIEmSJE3H9qMuQJtWVWuT3JPkIOCxwE3AocALu9cAu9ILLVfTCybHdO2P79rvAdYDF052nG7V5TSAebvvNYRPIkmSJA3GkLJ1+DhwMvCrwCeAFwB/XFUf6++U5DeAo4DDq+r+JFcCO3abf15V6yc7QFWtAFYAzF+4uGa5fkmSJGlgXu61dbgYeBG9FZQvdD+nJtkVIMk+SfYGFgA/6gLKk4Fnj6pgSZIkaXO5krIVqKpfJvky8ONuNeTyJE8BrksCcC/wKuAy4HVJbgHuAL42qpolSZKkzWVI2Qp0N8w/GzhuQ1tVfQj40ATdf2eifVTVrsOpTpIkSZpdXu7VuCRPBb4NXFFV3xp1PZIkSdKwuZLSuKr6OrDfqOuQJEmSthRXUiRJkiQ1xZAiSZIkqSmGFEmSJElNMaRIkiRJaoohRZIkSVJTDCmSJEmSmmJIkSRJktQUQ4okSZKkphhSJEmSJDXFJ85rIwfus4Cx5ctGXYYkSZLmKFdSJEmSJDXFkCJJkiSpKYYUSZIkSU0xpEiSJElqiiFFkiRJUlMMKZIkSZKaYkiRJEmS1BRDiiRJkqSmGFIkSZIkNcWQIkmSJKkphhRJkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSmGFEmSJElNSVWNugY1JslPgTtGXcdWYk/gh6MuYivifA3OuZoe52twztXgnKvpcb4G51w97IlVtdf4xu1HUYmad0dVLR11EVuDJGPO1eCcr8E5V9PjfA3OuRqcczU9ztfgnKupebmXJEmSpKYYUiRJkiQ1xZCiiawYdQFbEedqepyvwTlX0+N8Dc65GpxzNT3O1+Ccqyl447wkSZKkpriSIkmSJKkphpQ5JsmLktyR5NtJzpxge5L8ebf9liQHDzp2WzPDuVqbZE2S1UnGtmzlW94Ac/XkJNcl+UWSM6Yzdls0w/ny3Hrk9hO7//5uSfLVJM8cdOy2ZoZzNafOKxhovo7u5mp1krEkRww6dlszw7ny3Jq836FJ1ic5drpj54Sq8meO/ADzgO8A+wG/AtwMPHVcnxcD/wgEeDZw/aBjt6WfmcxVt20tsOeoP0dDc7U3cCjwfuCM6Yzd1n5mMl+eWxPO1XOAR3Wvf8f/Z01/rubaeTWN+dqVhy+LfwZwu+fW9ObKc2vy86Pr90/APwDHzsVza6ofV1LmlsOAb1fVd6vql8DfAEeP63M08Knq+RqwR5KFA47dlsxkruaaKeeqqn5QVSuBB6Y7dhs0k/maawaZq69W1Y+6t18Dfm3QsduYmczVXDTIfN1b3d8cgV2AGnTsNmYmczUXDXp+vBG4EPjBZoydEwwpc8s+wL/0vf/Xrm2QPoOM3ZbMZK6g9z/oy5OsSnLa0Kpsw0zOjbl2XsHMP7Pn1uReS291c3PGbu1mMlcwt84rGHC+khyT5HbgUuDU6YzdhsxkrsBza6P5SrIPcAzw0emOnUt84vzckgnaxv9rx2R9Bhm7LZnJXAE8t6r+LcnewBeT3F5VV89qhe2Yybkx184rmPln9tyaqGPyfHp/8d5wLfxcO7dmMlcwt84rGHC+qupi4OIkRwLvA44adOw2ZCZzBZ5bsPF8nQ38YVWtTx7Rfa6dW5vkSsrc8q/A4/ve/xrwbwP2GWTstmQmc0VVbfj9A+Bieku426qZnBtz7byCGX5mz62N5yrJM4CPA0dX1T3TGbsNmclczbXzCqZ5fnR/qd4/yZ7THbsNmMlceW5NPF9Lgb9JshY4FvhIkpcNOHbOMKTMLSuBxUn2TfIrwAnAJeP6XAK8Jj3PBtZV1fcGHLst2ey5SrJLkt0AkuwCvBC4dUsWv4XN5NyYa+cVzOAze25tPFdJngBcBLy6qr45nbHbmM2eqzl4XsFg8/WkdP/Mnd63N/4KcM8gY7cxmz1XnlsTz1dV7VtVi6pqEfB3wOur6nODjJ1LvNxrDqmqB5O8AfgCvW+Q+ERV3Zbkdd32j9L7lokXA98G7gdO2dTYEXyMLWImcwU8lt6SN/T+G/vrqrpsC3+ELWaQuUryq8AYsDvwUJK30PvGkp/MpfMKZjZfwJ54bo3/7/DdwGPo/UskwINVtdT/Zw0+V8yx/2fBwPP1Cnr/EPUA8DPg+O7mcM+tAecqiefWxPM1rbFbou4W+cR5SZIkSU3xci9JkiRJTTGkSJIkSWqKIUWSJElSUwwpkiRJkppiSJEkSZLUFEOKJEmSpKYYUiRJkiQ1xZAiSZIkqSn/P8v1eRo7HiWxAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot the feature importance\n",
    "plt.figure(figsize=(12, 8))\n",
    "plt.barh(feature_importance.index[:10], feature_importance.values[:10])\n",
    "plt.ylabel('Features')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The most important features are year of the car, odometer, description length, drive forward, cylinders, and fuel."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Mukc1qeEgpk4"
   },
   "source": [
    "# Question 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "id": "fb7PD1OzWljv"
   },
   "outputs": [],
   "source": [
    "# import function used to create clusters of data points normally distributed for classification\n",
    "from sklearn.datasets import make_classification\n",
    "\n",
    "# creating simulated data set\n",
    "X_sim, y_sim = make_classification(n_samples=1000, n_features=2,\n",
    "                           n_informative=2, n_redundant=0,\n",
    "                           random_state=0, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "jqSUgzuWXHmd",
    "outputId": "f3c5fb65-b88d-49ef-a05a-29094b2bd1fd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 3.41 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# (1) training error decreases as model 1/K increases\n",
    "# (2) test error first decreases and then increases as 1/K increases\n",
    "\n",
    "# create lists to store 1/K, training and testing errors\n",
    "model_flexibility = []\n",
    "training_error = []\n",
    "testing_error = []\n",
    "\n",
    "# creating a loop to modify the value of k to show\n",
    "for k in range(1, 101):  \n",
    "  # add 1/K to list\n",
    "  model_flexibility.append(1/k)\n",
    "\n",
    "  # creating KNN model with varying levels of k\n",
    "  knn_model = KNeighborsClassifier(n_neighbors=k)\n",
    "  \n",
    "  # fit model with our X and y\n",
    "  score = cross_validate(knn_model, X_sim, y_sim, n_jobs=-1, return_train_score=True)\n",
    "\n",
    "  # appending average error rate per K\n",
    "  training_error.append(1 - score['train_score'].mean())\n",
    "\n",
    "  # appending average error rate per K\n",
    "  testing_error.append(1 - score['test_score'].mean())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 388
    },
    "id": "gXnWMgeGheM2",
    "outputId": "87e88ba5-b0cc-40a3-bc1d-1ae26ba6add6"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting the scores\n",
    "plt.figure(figsize=(20, 6))\n",
    "plt.plot(model_flexibility, training_error, label = 'Training Error')\n",
    "plt.plot(model_flexibility, testing_error, label = 'Testing Error')\n",
    "\n",
    "plt.xlim((0,1))\n",
    "plt.xlabel('1/K (Model Flexibility)')\n",
    "plt.ylabel('Error Rate')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()\n",
    "# plt.xticks([0, 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Submission 2 - Test Data Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "# loading the test data set\n",
    "test_df = pd.read_csv('6Zz. final_predcomp_test_data (1).csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Redefining functiont to impute missing values\n",
    "def missing_value_impute(df):\n",
    "  # let's create a dummy variable column for missing values by iterating through all columns\n",
    "  # creating list to contain missing row indices\n",
    "  missing_row = []\n",
    "\n",
    "  # iterates through each row of large dataset\n",
    "  for index, row in df.iterrows():\n",
    "      \n",
    "      # checks each column to see if there are any missing values\n",
    "      missing_series = row.isnull()\n",
    "\n",
    "      # if any of the columns return true, add to our missing row list\n",
    "      if missing_series.any():\n",
    "          missing_row.append(index)\n",
    "  \n",
    "  # adding the dummy variable column\n",
    "  df.loc[missing_row, 'missing_value'] = 1\n",
    "\n",
    "  # fill the dummy variable columns w/ 0 for rows w/ non-missing data\n",
    "  df['missing_value'] = df['missing_value'].fillna(0)\n",
    "\n",
    "  # creating a list for categorical and numerical values\n",
    "  num_col = ['price', 'year', 'odometer', 'cylinders', 'description_length']\n",
    "\n",
    "  cat_col = ['transmission', 'drive', 'type', 'paint_color',\n",
    "            'size', 'title_status', 'condition', 'fuel',\n",
    "            'state', 'manufacturer', 'description_credit',\n",
    "             'description_owner', 'description_badcredit', 'missing_value']\n",
    "\n",
    "  # imputing the missing values using the mean for numerical columns and mode for categorical\n",
    "  # use for loop to loop through all columns\n",
    "  for col in df.columns[1:]:\n",
    "    if col in num_col:\n",
    "      df[col] = df[col].fillna(df[col].mean())\n",
    "    elif col in cat_col:\n",
    "      df[col] = df[col].fillna(df[col].mode().values[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "# use the created function to impute the missing values and create missing value dummy variable\n",
    "missing_value_impute(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "# construct dummy variable for below $12,000\n",
    "test_df['price_target'] = test_df['price'].apply(lambda x: 1 if x < 12000 else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "# list of categorical columns\n",
    "num_col = ['year', 'odometer', 'cylinders', 'description_length']\n",
    "\n",
    "cat_col = ['transmission', 'drive', 'type', 'paint_color',\n",
    "        'size', 'title_status', 'condition', 'fuel',\n",
    "        'state', 'manufacturer', 'description_credit', 'description_owner',\n",
    "        'description_badcredit', 'missing_value']\n",
    "\n",
    "# creating a function for dummy variables \n",
    "def get_dummies(df):\n",
    "    '''\n",
    "    This function takes the train/test dataset and creates dummy variables for the known categorical variables.\n",
    "    '''\n",
    "    # using pandas get dummies function to create dummy variables for each categorical item\n",
    "    dummy_df = pd.get_dummies(df[cat_col[:10]])\n",
    "\n",
    "    # saving dummy feature names to list\n",
    "    dummy_features = list(dummy_df.columns)\n",
    "\n",
    "    # merging original df w/ the new dummy variable columns\n",
    "    new_df = df.merge(dummy_df, left_index=True, right_index=True)\n",
    "\n",
    "    # return/output the new merged df and the dummy feature column names\n",
    "    return new_df, dummy_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "# using function to create new large df and the list of dummy column names\n",
    "new_test_df, dummy_col = get_dummies(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "# splitting the test data set into X and y\n",
    "# Set X and y\n",
    "X_testdata = new_test_df.loc[:, num_col + dummy_col]\n",
    "y_testdata = new_test_df['price_target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get unique list of columns from training and test data and then find the difference\n",
    "missing_cols = set(X.columns) - set(X_testdata.columns)\n",
    "\n",
    "# Add a missing column in test set with default value equal to 0\n",
    "for c in missing_cols:\n",
    "    X_testdata[c] = 0\n",
    "    \n",
    "# Ensure the order of column in the test set is in the same order as in training set\n",
    "X_testdata = X_testdata[X.columns]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test data accuracy: 0.9007333333333334\n"
     ]
    }
   ],
   "source": [
    "# predicting y^ for test data\n",
    "test_predictions = dt_model.predict(X_testdata)\n",
    "\n",
    "# returning accuracy score\n",
    "print(f'Test data accuracy: {dt_model.score(X_testdata,y_testdata)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Confusion Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "# importing confusion matrix functions from sklearn library\n",
    "from sklearn.metrics import confusion_matrix, plot_confusion_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot non-normalized (count statistics) confusion matrix\n",
    "disp = plot_confusion_matrix(dt_model, X_test, y_test,\n",
    "                             display_labels=['≥ $12,000','< $12,000'],\n",
    "                             cmap=plt.cm.Blues,\n",
    "                             normalize=None)\n",
    "disp.ax_.set_title(\"Confusion Matrix for Training Data (Without Normalization)\")\n",
    "\n",
    "# print(title)\n",
    "# print(disp.confusion_matrix)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot non-normalized (count statistics) confusion matrix\n",
    "disp = plot_confusion_matrix(dt_model, X_testdata, y_testdata,\n",
    "                             display_labels=['≥ $12,000','< $12,000'],\n",
    "                             cmap=plt.cm.Blues,\n",
    "                             normalize=None)\n",
    "disp.ax_.set_title(\"Confusion Matrix for Test Data (Without Normalization)\")\n",
    "\n",
    "# print(title)\n",
    "# print(disp.confusion_matrix)\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "Copy of Copy of Assignment3.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}