Cross Validation & Feature Scaling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing essential libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 865,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scipy as sp\n",
    "import os\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_selection import RFE\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import confusion_matrix\n",
    "#from imblearn.over_sampling import SMOTE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reading the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 866,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32561, 15)"
      ]
     },
     "execution_count": 866,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "income = pd.read_csv(\"income_data.csv\", sep=r'\\s*,\\s*', engine='python')\n",
    "income.head()\n",
    "#QC:Checking number of rows and columns \n",
    "type(income)\n",
    "income.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 867,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age                  0\n",
       "workclass         1836\n",
       "fnlwgt               0\n",
       "education            0\n",
       "education-num        0\n",
       "marital-status       0\n",
       "occupation        1843\n",
       "relationship         0\n",
       "race                 0\n",
       "sex                  0\n",
       "capital-gain         0\n",
       "capital-loss         0\n",
       "hours-per-week       0\n",
       "native-country     583\n",
       "income               0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 867,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Checking for null values\n",
    "income.isnull()\n",
    "#Replacing missing values with numpy NaN values  \n",
    "income.replace('', np.nan, inplace=True)\n",
    "income.replace('?', np.nan, inplace=True)\n",
    "#Columns with null/NaN values\n",
    "income.isnull().any()\n",
    "#summary of missing values\n",
    "income.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 868,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>fnlwgt</th>\n",
       "      <th>education</th>\n",
       "      <th>education-num</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>capital-gain</th>\n",
       "      <th>capital-loss</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>77516</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>2174</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>83311</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>215646</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Divorced</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53</td>\n",
       "      <td>Private</td>\n",
       "      <td>234721</td>\n",
       "      <td>11th</td>\n",
       "      <td>7</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>338409</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>Wife</td>\n",
       "      <td>Black</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>Cuba</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age         workclass  fnlwgt  education  education-num  \\\n",
       "0   39         State-gov   77516  Bachelors             13   \n",
       "1   50  Self-emp-not-inc   83311  Bachelors             13   \n",
       "2   38           Private  215646    HS-grad              9   \n",
       "3   53           Private  234721       11th              7   \n",
       "4   28           Private  338409  Bachelors             13   \n",
       "\n",
       "       marital-status         occupation   relationship   race     sex  \\\n",
       "0       Never-married       Adm-clerical  Not-in-family  White    Male   \n",
       "1  Married-civ-spouse    Exec-managerial        Husband  White    Male   \n",
       "2            Divorced  Handlers-cleaners  Not-in-family  White    Male   \n",
       "3  Married-civ-spouse  Handlers-cleaners        Husband  Black    Male   \n",
       "4  Married-civ-spouse     Prof-specialty           Wife  Black  Female   \n",
       "\n",
       "   capital-gain  capital-loss  hours-per-week native-country income  \n",
       "0          2174             0              40  United-States  <=50K  \n",
       "1             0             0              13  United-States  <=50K  \n",
       "2             0             0              40  United-States  <=50K  \n",
       "3             0             0              40  United-States  <=50K  \n",
       "4             0             0              40           Cuba  <=50K  "
      ]
     },
     "execution_count": 868,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#trim whitespaces if any for columns of type string\n",
    "income_trimmed = income.apply(lambda x: x.str.strip() if x.dtype == \"object\" else x)\n",
    "income_trimmed.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 869,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>fnlwgt</th>\n",
       "      <th>education-num</th>\n",
       "      <th>capital-gain</th>\n",
       "      <th>capital-loss</th>\n",
       "      <th>hours-per-week</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>32561.000000</td>\n",
       "      <td>3.256100e+04</td>\n",
       "      <td>32561.000000</td>\n",
       "      <td>32561.000000</td>\n",
       "      <td>32561.000000</td>\n",
       "      <td>32561.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>38.581647</td>\n",
       "      <td>1.897784e+05</td>\n",
       "      <td>10.080679</td>\n",
       "      <td>1077.648844</td>\n",
       "      <td>87.303830</td>\n",
       "      <td>40.437456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>13.640433</td>\n",
       "      <td>1.055500e+05</td>\n",
       "      <td>2.572720</td>\n",
       "      <td>7385.292085</td>\n",
       "      <td>402.960219</td>\n",
       "      <td>12.347429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>17.000000</td>\n",
       "      <td>1.228500e+04</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>28.000000</td>\n",
       "      <td>1.178270e+05</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>37.000000</td>\n",
       "      <td>1.783560e+05</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>48.000000</td>\n",
       "      <td>2.370510e+05</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>45.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>90.000000</td>\n",
       "      <td>1.484705e+06</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>99999.000000</td>\n",
       "      <td>4356.000000</td>\n",
       "      <td>99.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                age        fnlwgt  education-num  capital-gain  capital-loss  \\\n",
       "count  32561.000000  3.256100e+04   32561.000000  32561.000000  32561.000000   \n",
       "mean      38.581647  1.897784e+05      10.080679   1077.648844     87.303830   \n",
       "std       13.640433  1.055500e+05       2.572720   7385.292085    402.960219   \n",
       "min       17.000000  1.228500e+04       1.000000      0.000000      0.000000   \n",
       "25%       28.000000  1.178270e+05       9.000000      0.000000      0.000000   \n",
       "50%       37.000000  1.783560e+05      10.000000      0.000000      0.000000   \n",
       "75%       48.000000  2.370510e+05      12.000000      0.000000      0.000000   \n",
       "max       90.000000  1.484705e+06      16.000000  99999.000000   4356.000000   \n",
       "\n",
       "       hours-per-week  \n",
       "count    32561.000000  \n",
       "mean        40.437456  \n",
       "std         12.347429  \n",
       "min          1.000000  \n",
       "25%         40.000000  \n",
       "50%         40.000000  \n",
       "75%         45.000000  \n",
       "max         99.000000  "
      ]
     },
     "execution_count": 869,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "income.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 870,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32537, 15)"
      ]
     },
     "execution_count": 870,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Looking for duplicates in the data\n",
    "income_trimmed.duplicated().sum()\n",
    "#Removing duplicate entries from our data\n",
    "income_c =income_trimmed.drop_duplicates()\n",
    "income_c.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Pre-Processing & EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 871,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age               0\n",
       "workclass         0\n",
       "fnlwgt            0\n",
       "education         0\n",
       "education-num     0\n",
       "marital-status    0\n",
       "occupation        0\n",
       "relationship      0\n",
       "race              0\n",
       "sex               0\n",
       "capital-gain      0\n",
       "capital-loss      0\n",
       "hours-per-week    0\n",
       "native-country    0\n",
       "income            0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 871,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Handling Nulls by replacing them with most frequent occurrence\n",
    "#~6% of the rows collectively have incomplete data \n",
    "index=income_c.index\n",
    "\n",
    "from sklearn.impute import SimpleImputer as Imputer\n",
    "#creating an object for SimpleImputer\n",
    "imp_freq=Imputer(missing_values=np.nan,strategy='most_frequent')\n",
    "income_c=pd.DataFrame(imp_freq.fit_transform(income_c))\n",
    "income_c.columns = income.columns\n",
    "income_c.index = index\n",
    "\n",
    "#Check if imputation worked\n",
    "income_c.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 872,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Bachelors', 'HS-grad', 'Not-HS-Grad', 'Masters', 'Some-college',\n",
       "       'Assoc-acdm', 'Assoc-voc', 'Doctorate', 'Prof-school'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 872,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#merging levels of education\n",
    "income_c['education']=np.where(income_c['education'] =='11th', 'Not-HS-Grad', income_c['education'])\n",
    "income_c['education']=np.where(income_c['education'] =='9th', 'Not-HS-Grad', income_c['education'])\n",
    "income_c['education']=np.where(income_c['education'] =='7th-8th', 'Not-HS-Grad', income_c['education'])\n",
    "income_c['education']=np.where(income_c['education'] =='5th-6th', 'Not-HS-Grad', income_c['education'])\n",
    "income_c['education']=np.where(income_c['education'] =='10th', 'Not-HS-Grad', income_c['education'])\n",
    "income_c['education']=np.where(income_c['education'] =='1st-4th', 'Not-HS-Grad', income_c['education'])\n",
    "income_c['education']=np.where(income_c['education'] =='Preschool', 'Not-HS-Grad', income_c['education'])\n",
    "income_c['education']=np.where(income_c['education'] =='12th', 'Not-HS-Grad', income_c['education'])\n",
    "\n",
    "#unique levels of education\n",
    "income_c['education'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 873,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([6, 2, 1, 7, 3, 5, 4, 9, 8], dtype=int64)"
      ]
     },
     "execution_count": 873,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#modify education_num in proportion to education\n",
    "# Create mapper\n",
    "scale_mapper = {'Not-HS-Grad':1, \n",
    "                'HS-grad':2,\n",
    "                'Some-college':3,\n",
    "               'Assoc-voc':4,\n",
    "               'Assoc-acdm':5,\n",
    "               'Bachelors':6,\n",
    "               'Masters':7,\n",
    "               'Prof-school':8,\n",
    "               'Doctorate':9}\n",
    "income_c['education']=income_c['education'].replace(scale_mapper)\n",
    "income_c['education'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 874,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "education  education-num\n",
       "1          1                   50\n",
       "           2                  166\n",
       "           3                  332\n",
       "           4                  645\n",
       "           5                  514\n",
       "           6                  933\n",
       "           7                 1175\n",
       "           8                  433\n",
       "2          9                10494\n",
       "3          10                7282\n",
       "4          11                1382\n",
       "5          12                1067\n",
       "6          13                5353\n",
       "7          14                1722\n",
       "8          15                 576\n",
       "9          16                 413\n",
       "dtype: int64"
      ]
     },
     "execution_count": 874,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#1:1 mapping between columns 'education-num' and 'education' in the original dataset\n",
    "#education_num being a quantitative variable will be more valuable in our model than the categorical variable as it has order \n",
    "#based on education qualification\n",
    "income_c.groupby(['education','education-num']).size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 875,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '# of records')"
      ]
     },
     "execution_count": 875,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "pd.crosstab(income_c['relationship'],income_fact['income']).plot(kind='bar')\n",
    "plt.title('Income against Education')\n",
    "plt.xlabel('Education')\n",
    "plt.ylabel('# of records')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 876,
   "metadata": {},
   "outputs": [],
   "source": [
    "del income_c['education-num']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 877,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'High Salary Proportion')"
      ]
     },
     "execution_count": 877,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAFoCAYAAABAGRzjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsnXe4XFXVh98foYSShBYxEEIiRDoJGAKKKCAGCBBQaVLlQ1EBKaKIFaQo+Kl8UgQRkNCEgCgRKUGq1JBA6CWRGhIh0kQQJPD7/th7ksnN3DJ39tybcNf7PPPMnH3OWXvPuXPPOnvtVWSbIAiCIOgoi3T3AIIgCIKFi1AcQRAEQV2E4giCIAjqIhRHEARBUBehOIIgCIK6CMURBEEQ1EUojiDoIiTtJWlCd4+jGknHSrqokKyzJP2wK/oKupdQHEFxJD0jaevuHseChu2LbY9qVI4kS1qjjf1bSJreaD8tZA6S9O+qlyW9WbW9ue2v2T6+WWMIFhwW7e4BBEGw4GP7OWCZyrYkA8NsT+u+UQXdRcw4gqYi6UuSbpf0c0mvSnpa0nZV+5eX9DtJM/L+P1Xt+4qkaZJekTRe0spV+yzpIElTJb0h6XhJq0u6S9K/JI2TtHjV8TtImiLpNUl3StqgjTH/StLzWc5kSZtX7VtS0tg81sckHVX9ZC3paEl/z2N6VNLnWl6LFt/ha/k7vCrpDEnK+9aQdKuk1yX9U9Jluf22fPoD+Ul/9w78DVbO1++VfD2/0uKQ3pIuy2O+T9Kw9mS20s/5kk6QtDRwLbBy1Yxk5RrHb5r/Fq9JekDSFp3pN+h6QnEEXcEmwBPAisDPgHMrN0jgQmApYF3gQ8ApAJK2An4K7AYMAJ4FLm0hd1vgY8CmwFHA2cBewKrAesAXs6yNgPOArwIrAL8BxktaopXx3gsMB5YHLgEul9Q77zsGGAx8BPgssHeLc/8ObA70A34MXCRpQBvXZgdgY2BY/q7b5PbjgQnAcsBA4DQA25/K+4fZXsb2ZW3IrvB7YDqwMrAL8BNJn6navxNwedX3/ZOkxTogtya23wS2A2bkMS5je0b1MZJWAf4CnJD7/RbwB0n9O9tv0HWE4gi6gmdt/9b2e8BYkiJYKd9QtwO+ZvtV2+/avjWfsxdwnu37bL8DfBf4uKTBVXJPtv0v248ADwMTbD9l+3XSE++G+bivAL+xfY/t92yPBd4hKZz5sH2R7Zdtz7b9C2AJYM28ezfgJ3m804FTW5x7ue0Ztt/PN/WpwMg2rs1Jtl/LpqCbSQoL4F1gNWBl22/bvr1VCW0gaVXgk8B3spwpwDnAPlWHTbZ9he13gV8CvWnl2hRkb+Aa29fka3UDMAkY3eR+gwKE4gi6gn9UPth+K39chjQzeMX2qzXOWZk0y6ic92/gZWCVqmNerPr8nxrbFZv8asCR2STymqTXct/zmU8AJB2ZzVCv52P7kWZLlXE9X3X48y3O3bfKJPYaaeazIq3zj6rPb1WN+ShAwERJj0j6nzZktMXKpGv8RlXbs8x7Hed8B9vvM3d20kxWA3Zt8Tf5JOmhIljAicXxoDt5Hlhe0rK2X2uxbwbp5gJAtpuvALzQyX5OtH1iewfm9YzvAJ8BHrH9vqRXSTdxgJkk09GjeXvVqnNXA36bz73L9nuSplSd22Fs/4M0U0LSJ4G/SrqtE4vRM0jXuE+V8hjEvNex+jssQvp+85iWOkF7abefBy603XK9JVgIiBlH0G3YnkkyKf1a0nKSFpNUseFfAuwvaXhei/gJcI/tZzrR1W+Br0naRImlJW0vqU+NY/sAs4FZwKKSfgT0rdo/DvhuHu8qwCFV+5Ym3TBnAUjanzTjqBtJu0oamDdfzXLfy9svktZY2sX288CdwE8l9c5OAQcAF1cd9jFJn5e0KHA4yYx3d2fGXcWLwAqS+rWy/yJgR0nbSOqVx7ZF1XcOFmBCcQTdzT4ke/7jwEukGxe2bwR+CPyB9JS/OrBHZzqwPYn09H466SY8DfhSK4dfT1JmT5JMOm8zrznqOJIp52ngr8AVpBstth8FfgHcRbpxrg/c0ZkxkxbM75H0b2A8cJjtp/O+Y4Gx2cSzWyvnVz/xf5G0oD8D+CNwTF5TqHAVsDvp2uwDfD6vd3Qa24+TFuWfyuNcucX+50mL8t8jKdrngW8T96SFAkUhpyDoPJK+Duxh+9PdPZYKksYAx9ke3u7BQdAJQrsHQR1IGiBpM0mLSFoTOJL0FL9AkM1NXyB5KAVBU4jF8SCoj8VJcSBDgNdIsSW/7tYRZfJ6wvPAZGDfbh5O8AEmTFVBEARBXYSpKgiCIKiLUBxBEARBXXwg1zhWXHFFDx48uLuHEQRBsFAxefLkf9puN1/YB1JxDB48mEmTwqkkCIKgHiQ92/5RYaoKgiAI6qTpiiOnE7hf0tV5e4ike3INgsuUayZIWiJvT8v7B1fJ+G5uf0LSNrV7CoIgCLqCrphxHAY8VrV9MnCK7aGkFAcH5PYDgFdtr0GqyXAygKR1SKkm1iXVX/i1pF5dMO4gCIKgBk1d48gJy7YHTgS+mYv3bAXsmQ8ZS8q7cyYpb82xuf0K4PR8/E7Apbkmw9OSppHqG9zVzLEHQbDw8+677zJ9+nTefvvt7h7KAkXv3r0ZOHAgiy3WuXpdzV4c/z9SXYFKFtIVgNdsz87b05lbF2AVcjI527MlvZ6PX4V5M3VWnzMHSQcCBwIMGjSo7LcIgmChZPr06fTp04fBgwczt+hkz8Y2L7/8MtOnT2fIkCGdktE0U5WkHYCXbE+ubq5xqNvZ19Y5cxvss22PsD2if/+oPhkEAbz99tussMIKoTSqkMQKK6zQ0CysmTOOzYAxkkaTSlH2Jc1AlpW0aJ51VBeMmU4qKDM9J2rrB7xS1V6hRJGZIAh6CKE05qfRa9K0GYft79oeaHswaXH7Jtt7keoq75IP249UCwBSzYH98udd8vHO7Xtkr6shwFBgYrPGHQRB0JJPfOIT3T2EBYruCAD8DnCppBOA+4Fzc/u5wIV58fsVctEe249IGkcq1TkbONj2e/OLDYKgOxl89F9qtj9z0vZdPJLy3Hnnnd09hAWKLgkAtH2L7R3y56dsj7S9hu1ds7cUtt/O22vk/U9VnX+i7dVtr2n72q4YcxAEQYVlllkGgFtuuYUtttiCXXbZhbXWWou99tqLSobxe++9l0984hMMGzaMkSNH8sYbb/D222+z//77s/7667Phhhty8803A3D++eez8847s+OOOzJkyBBOP/10fvnLX7Lhhhuy6aab8sorrwDw97//nW233ZaPfexjbL755jz++OPdcwFa8IFMORIEQdAs7r//fh555BFWXnllNttsM+644w5GjhzJ7rvvzmWXXcbGG2/Mv/71L5Zcckl+9atfAfDQQw/x+OOPM2rUKJ588kkAHn74Ye6//37efvtt1lhjDU4++WTuv/9+jjjiCC644AIOP/xwDjzwQM466yyGDh3KPffcw0EHHcRNN93UnV8fCMURBEFQFyNHjmTgwIEADB8+nGeeeYZ+/foxYMAANt54YwD69u0LwO233843vvENANZaay1WW221OYpjyy23pE+fPvTp04d+/fqx4447ArD++uvz4IMP8u9//5s777yTXXfddU7f77zzTpd9z7YIxREEQVAHSyyxxJzPvXr1Yvbs2diu6anUVqG8ajmLLLLInO1FFlmE2bNn8/7777PssssyZcqUgqMvQyQ5DIIgaJC11lqLGTNmcO+99wLwxhtvMHv2bD71qU9x8cUXA/Dkk0/y3HPPseaaa3ZIZt++fRkyZAiXX345kJTQAw880JwvUCehOIIgCBpk8cUX57LLLuMb3/gGw4YN47Of/Sxvv/02Bx10EO+99x7rr78+u+++O+eff/48M432uPjiizn33HMZNmwY6667LldddVX7J3UBH8ia4yNGjHDU4wiCrmVBdMd97LHHWHvttbut/wWZWtdG0mTbI9o7N2YcQRAEQV2E4giCIAjqIhRHEARBUBehOIIgCIK6CMURBEEQ1EUojiAIgqAuQnEEQRAsoNxyyy3069eP4cOHM3z4cI477rg5+6677jrWXHNN1lhjDU466aQ57VtssQWVcIRnnnmGoUOHcv311xcdV6QcCYKgx9BarEln6UyMyn//+1/effddll566Q4dv/nmm3P11VfP0/bee+9x8MEHc8MNNzBw4EA23nhjxowZwzrrrDPnmOnTp7PNNtvwi1/8gm222abucbZFzDiCIAi6gMcee4wjjzySNddcc06iw84yceJE1lhjDT7ykY+w+OKLs8cee8wTVf6Pf/yDUaNGccIJJzBmzJhGhz4fMeMIgiBoEm+++Sbjxo3j3HPPxTb7778/Dz74IH369AHgiCOOmFOjo5o99tiDo48+GoC77rqLYcOGsfLKK/Pzn/+cddddlxdeeIFVV51bUXvgwIHcc889c7b33XdfTjjhhHky65YkFEcQBEGTGDBgABtssAHnnHMOa6211nz7TznllDbP32ijjXj22WdZZplluOaaa9h5552ZOnVqzay71dl5t956ay688EK+9KUvsdRSSzX+RVrQNFOVpN6SJkp6QNIjkn6c28+X9LSkKfk1PLdL0qmSpkl6UNJGVbL2kzQ1v/Zrrc8gCIIFiSuuuIJVVlmFz33ucxx33HE8++yz8+w/4ogj5ix8V78qi919+/adU31w9OjRvPvuu/zzn/9k4MCBPP/883PkTJ8+nZVXXnnO9lFHHcUmm2zCrrvuyuzZs4t/r2bOON4BtrL9b0mLAbdLqpR9/bbtK1ocvx0wNL82Ac4ENpG0PHAMMAIwMFnSeNuvNnHsQRAEDTNq1ChGjRrFyy+/zEUXXcROO+3EiiuuyDnnnMPgwYPbnXH84x//YKWVVkISEydO5P3332eFFVZg2WWXZerUqTz99NOsssoqXHrppVxyySXznHvKKaew5557csABB3D++efXrBfSWZo243Di33lzsfxqKxXvTsAF+by7gWUlDQC2AW6w/UpWFjcA2zZr3EEQBKVZYYUVOOyww5gyZQo/+clP6NWrV4fOu+KKK1hvvfUYNmwYhx56KJdeeimSWHTRRTn99NPZZpttWHvttdltt91Yd9115zlXEmPHjmXmzJkcddRRRb9PU9c4JPUCJgNrAGfYvkfS14ETJf0IuBE42vY7wCrA81WnT89trbUHQRDURXemeK8wcuTIDh97yCGHcMghh9TcN3r0aEaPHj1f+y233DLn8+KLL86ECRPqHmN7NNUd1/Z7tocDA4GRktYDvgusBWwMLA98Jx9eax7lNtrnQdKBkiZJmjRr1qwi4w+CIAjmp0viOGy/BtwCbGt7ZjZHvQP8Dqio3+nAqlWnDQRmtNHeso+zbY+wPaJ///5N+BZBEAQBNNerqr+kZfPnJYGtgcfzugVKKzU7Aw/nU8YD+2bvqk2B123PBK4HRklaTtJywKjcFgRBEHQDzVzjGACMzesciwDjbF8t6SZJ/UkmqCnA1/Lx1wCjgWnAW8D+ALZfkXQ8cG8+7jjbrzRx3EEQBEEbNE1x2H4Q2LBG+1atHG/g4Fb2nQecV3SAQRAEQaeIXFVBEARBXYTiCIIg6Ca+9KUvMWTIkDkR41OmTAHANoceeihrrLEGG2ywAffddx+Q0qSvt956c87/7W9/y0YbbcSrr3ZtPHTkqgqCoOdwbL/C8l5vc/err77Kcsst1+Yx//u//8suu+wyT9u1117L1KlTmTp1Kvfccw9f//rX50liCHDhhRdy2mmncdNNN7XbR2lixhEEQdAkRowYwZ577slNN91UMzFha1x11VXsu+++SGLTTTfltddeY+bMmXP2jxs3jpNOOokJEyaw4oorNmPobRKKIwiCoEk8+eST7Lnnnpx++umss846/OQnP2HGjHnD0L7//e+zwQYbcMQRR/DOO+8A1Eyb/sILLwDw7LPPcsghhzBhwgQ+/OEPd92XqSIURxAEQZPo1asXO+ywA1deeSW33XYbTz31FIMGDWLixIkA/PSnP+Xxxx/n3nvv5ZVXXuHkk08GaDNtev/+/Rk0aBDjxo3rui/SglAcQRAETeT111/n7LPPZsyYMTz55JOce+65bLDBBkCq1yGJJZZYgv3333+OQmkrbfpSSy3Ftddey1lnncXFF1/c9V+IHrY43lq94QUh8VkQBB889t57b+666y523XVXLrjgAoYOHTrP/pkzZzJgwABs86c//WmOx9SYMWM4/fTT2WOPPbjnnnvo168fAwYM4JlnngHSrOO6665jiy22YMUVVyxeU7w9epTiCIIg6Ep22203zj//fBZdtPatdq+99mLWrFnYZvjw4Zx11llAynx7zTXXsMYaa7DUUkvxu9/9br5zhwwZwvjx4xk9ejRXXnklm2yySVO/SzWhOIIg6Dm04z5bmjFjxrS5/6abbqrZLokzzjhjvvbBgwfz8MMPz9keNmzYnEXzriTWOIIgCIK6CMURBEEQ1EWYqhYiYnE/CIIFgVAcQbCAEA8GzcH2nBiIIFFPFHstwlQVBMEHlt69e/Pyyy83fKP8IGGbl19+md69e3daRsw4giD4wDJw4ECmT5/OrFmzunsoCxS9e/dm4MCBnT4/FEcQBB9YFltsMYYMGdLdw/jAEaaqIAiCoC6apjgk9ZY0UdIDkh6R9OPcPkTSPZKmSrpM0uK5fYm8PS3vH1wl67u5/QlJXRtbHwRBEMxDM2cc7wBb2R4GDAe2lbQpcDJwiu2hwKvAAfn4A4BXba8BnJKPQ9I6wB7AusC2wK8l9WriuIMgCII2aJricOLfeXOx/DKwFXBFbh8L7Jw/75S3yfs/o+RDtxNwqe13bD8NTANGNmvcQRAEQds0dY1DUi9JU4CXgBuAvwOv2Z6dD5kOrJI/rwI8D5D3vw6sUN1e45zqvg6UNEnSpPCgCIIgaB5NVRy237M9HBhImiWsXeuw/F4rQsdttLfs62zbI2yP6N+/f2eHHARBELRDl3hV2X4NuAXYFFhWUsUNeCBQqaM4HVgVIO/vB7xS3V7jnCAIgqCLaaZXVX9Jy+bPSwJbA48BNwO75MP2A67Kn8fnbfL+m5zCPccDe2SvqyHAUGBis8YdBEEQtE0zAwAHAGOzB9QiwDjbV0t6FLhU0gnA/cC5+fhzgQslTSPNNPYAsP2IpHHAo8Bs4GDb7zVx3EEQBEEbNE1x2H4Q2LBG+1PU8Iqy/TawayuyTgROLD3GIAiCoH4icjwIgiCoi1AcQRAEQV2E4giCIAjqIhRHEARBUBehOIIgCIK6CMURBEEQ1EUojiAIgqAuQnEEQRAEdRGKIwiCIKiLUBxBEARBXYTiCIIgCOoiFEcQBEFQF83MjhsEQbBQMPjov9Rsf+ak7bt4JAsHMeMIgiAI6iIURxAEQVAXoTiCIAiCumhXcUharysGEgRBECwcdGTGcZakiZIOqtQQ7wiSVpV0s6THJD0i6bDcfqykFyRNya/RVed8V9I0SU9I2qaqfdvcNk3S0XV9wyAIgqAo7XpV2f6kpKHA/wCTJE0Efmf7hnZOnQ0cafs+SX2AyZIq55xi++fVB0tah1RnfF1gZeCvkj6ad58BfBaYDtwrabztRzv4HYMgCIKCdMgd1/ZUST8AJgGnAhtKEvA921e2cs5MYGb+/Iakx4BV2uhmJ+BS2+8AT0uaxtza5NNyrXIkXZqPDcURBEHQDXRkjWMDSacAjwFbATvaXjt/PqUjnUgaDGwI3JObDpH0oKTzJC2X21YBnq86bXpua629ZR8HSpokadKsWbM6MqwgCIKgE3RkjeN04D5gmO2Dbd8HYHsG8IP2Tpa0DPAH4HDb/wLOBFYHhpNmJL+oHFrjdLfRPm+DfbbtEbZH9O/fv/1vFQRBEHSKNk1VknoBz9u+sNb+1tqrzl+MpDQurpi0bL9Ytf+3wNV5czqwatXpA4EZ+XNr7UEQBEEX0+aMw/Z7wAqSFq9XcF4DORd4zPYvq9oHVB32OeDh/Hk8sIekJSQNAYYCE4F7gaGShuRx7JGPDYIgCLqBjiyOPwvcIWk88GalsVoZtMJmwD7AQ5Km5LbvAV+UNJxkbnoG+GqW94ikcaRF79nAwVlxIekQ4HqgF3Ce7Uc69vWCIAiC0nREcczIr0WAPh0VbPt2aq9PXNPGOScCJ9Zov6at84IgCIKuoyNxHD/uioEEQRAECwftKg5J/YGjSIF5vSvttrdq4riCIAiCBZSOuONeDDwODAF+TFqXuLeJYwqCIAgWYDqiOFawfS7wru1bbf8PsGmTxxUEQRAsoHRkcfzd/D5T0vakhfKBzRtSEARBsCDTEcVxgqR+wJHAaUBf4IimjioIgiBYYOmIV1Ulsvt1YMvmDicIgiBY0GlVcUg6jRo5oSrYPrQpIwqCIAgWaNqacUzqslEEQRAECw2tKg7bY7tyIEEQBMHCQUcDAL8DrEMEAAZBEPR4OhoA+BgRABgEQRAQAYBBEARBnUQAYBAEQVAXEQAYBEEQ1EUEAAZBEAR10VYA4FeAW2xPzWVgzwM+T6oIuJ/t+7tojEEXMfjov9Rsf+ak7bt4JEEQLMi0tTh+GMmDCuCLwAbAR4BvAqe2J1jSqpJulvSYpEckHZbbl5d0g6Sp+X253C5Jp0qaJulBSRtVydovHz9V0n6d+6pBEARBCdoyVc22XVkY3wG4wPbLwF8l/awDsmcDR9q+T1IfYLKkG4AvATfaPknS0cDRpDiR7YCh+bUJcCawiaTlgWOAEaQUKJMljbf9ar1fNgiCbuDYfq20v9614wiK0daM431JAyT1Bj4D/LVq35LtCbY90/Z9+fMbpFiQVYCdgEpU+lhg5/x5J5Jysu27gWUlDQC2AW6w/UpWFjcA23b4GwZBEARFaWvG8SNSvqpewHjbjwBI+jTwVD2dSBoMbAjcA6xkeyYk5SLpQ/mwVYDnq06bnttaaw+CIAi6gbZyVV0taTWgTwuz0CRg9452IGkZ4A/A4bb/ldbZax9aaxhttLfs50DgQIBBgwZ1dHhBEARBnbTpjmt7NvBqi7Y3Oypc0mIkpXGx7Stz84uSBuTZxgDgpdw+HVi16vSBpGDD6cAWLdpvqTHWs4GzAUaMGNFqOvigE4SNOgiCKjqScqRTZBfec4HHbP+yatd4oOIZtR9wVVX7vtm7alPg9WzSuh4YJWm57IE1KrcFQRAE3UBHIsc7y2bAPsBDkqbktu8BJwHjJB0APAfsmvddA4wGpgFvAfsD2H5F0vHMTax4nO1XmjjuIAiCoA06pDgkrQKsVn287dvaOsf27dRen4DkpdXyeAMHtyLrPFIAYhAEQdDNdKQex8mkxfBHgfdys4E2FUcQBEHwwaQjM46dgTVtv9PswQRBEAQLPh1ZHH8KWKzZAwmCIAgWDtpKcngayST1FjBF0o3AnFmH7UObP7wgCIJgQaMtU9Wk/D6Z5CobBEEQBG1Gjo9tbV8QBEHQc+mIV9VDzJ/i43XSjOSEnDE3CIIg6CF0xKvqWpIb7iV5ew9SfMbrwPnAjk0ZWRAEQbBA0hHFsZntzaq2H5J0h+3NJO3drIEFwYJGVEgMgkRH3HGXkbRJZUPSSGCZvDm7KaMKgiAIFlg6MuP4MnBeTo8u4F/AlyUtDfy0mYMLgiAIFjzaVRy27wXWl9QPkO3XqnaPa9rIgiAIggWStgIA97Z9kaRvtmgHoEWq9CAIgqCH0NaMY+n83qcrBhIEQRAsHLQVAPib/P7jrhtOEARBsKDTlqnq1LZOjFxVQRAEPZO2TFWTqz7/GDimyWMJgiAIFgI6lKtK0uGRuyoIgiCAjgUAwvy5qtpF0nmSXpL0cFXbsZJekDQlv0ZX7fuupGmSnpC0TVX7trltmqSj6x1HEARBUJaOKo7OcD6wbY32U2wPz69rACStQ8qBtW4+59eSeknqBZwBbAesA3wxHxsEQRB0E20tjr/B3JnGUpL+VdkF2HbftgTbvk3S4A6OYyfg0lye9mlJ04CRed8020/lMV2aj320g3KDIAiCwrQ647Ddx3bf/Fq06nOf9pRGOxwi6cFsylout60CPF91zPTc1lr7fEg6UNIkSZNmzZrVwPCCIAiCtmimqaoWZwKrA8OBmcAvcrtqHOs22udvtM+2PcL2iP79+5cYaxAEQVCDjiQ5LIbtFyufJf0WuDpvTgdWrTp0IDAjf26tPQiCIOgGunTGIWlA1ebngIrH1XhgD0lLSBoCDAUmAvcCQyUNkbQ4aQE96p8HQRB0I02bcUj6PbAFsKKk6aQAwi0kDSeZm54Bvgpg+xFJ40iL3rOBg22/l+UcAlwP9ALOs/1Is8YcBEEQtE/TFIftL9ZoPreN408ETqzRfg1wTcGhBUEQBA3Q1YvjQRAEwUJOKI4gCIKgLkJxBEEQBHURiiMIgiCoi1AcQRAEQV2E4giCIAjqoksjx4OgmQw++i812585afsuHkkQdC1d/duPGUcQBEFQF6E4giAIgroIxREEQRDURSiOIAiCoC5CcQRBEAR1EYojCIIgqItQHEEQBEFdhOIIgiAI6iIURxAEQVAXoTiCIAiCumia4pB0nqSXJD1c1ba8pBskTc3vy+V2STpV0jRJD0raqOqc/fLxUyXt16zxBkEQBB2jmTOO84FtW7QdDdxoeyhwY94G2A4Yml8HAmdCUjSkWuWbACOBYyrKJgiCIOgemqY4bN8GvNKieSdgbP48Fti5qv0CJ+4GlpU0ANgGuMH2K7ZfBW5gfmUUBEEQdCFdvcaxku2ZAPn9Q7l9FeD5quOm57bW2udD0oGSJkmaNGvWrOIDD4IgCBILSlp11WhzG+3zN9pnA2cDjBgxouYxQRAEdXFsv1baX+/acSxgdPWM48VsgiK/v5TbpwOrVh03EJjRRnsQBEHQTXS14hgPVDyj9gOuqmrfN3tXbQq8nk1Z1wOjJC2XF8VH5bYgCIKgm2iaqUrS74EtgBUlTSd5R50EjJN0APAcsGs+/BpgNDANeAvYH8D2K5KOB+7Nxx1nu+WCexDT6SAIupCmKQ7bX2xl12dqHGvg4FbknAecV3BoQRAEQQNE5HgQBEFQFwuKV1XQAxh89F9qtj9z0vY+t5ayAAAgAElEQVRdPJIgCBohZhxBEARBXYTiCIIgCOoiFEcQBEFQF6E4giAIgroIxREEQRDURSiOIAiCoC5CcQRBEAR1EYojCIIgqItQHEEQBEFdhOIIgiAI6iIURxAEQVAXkasqCIKgyXzQ8rTFjCMIgiCoi1AcQRAEQV2E4giCIAjqolsUh6RnJD0kaYqkSblteUk3SJqa35fL7ZJ0qqRpkh6UtFF3jDkIgiBIdOeMY0vbw22PyNtHAzfaHgrcmLcBtgOG5teBwJldPtIgCIJgDguSqWonYGz+PBbYuar9AifuBpaVNKA7BhgEQRB0nzuugQmSDPzG9tnASrZnAtieKelD+dhVgOerzp2e22YWG82x/Vppf71YF0EQBB8UuktxbGZ7RlYON0h6vI1jVaPN8x0kHUgyZTFo0KAyowyCIAjmo1tMVbZn5PeXgD8CI4EXKyao/P5SPnw6sGrV6QOBGTVknm17hO0R/fv3b+bwgyAIejRdrjgkLS2pT+UzMAp4GBgP7JcP2w+4Kn8eD+ybvas2BV6vmLSCIAiCrqc7TFUrAX+UVOn/EtvXSboXGCfpAOA5YNd8/DXAaGAa8Bawf9cPOQiC7uaDlrZjYabLFYftp4BhNdpfBj5To93AwV0wtCAIgqADLEjuuEEQBMFCQCiOIAiCoC5CcQRBEAR1EYojCIIgqIso5BQEQfBBpUlZMWLGEQRBENRFzDiCIAi6i4U0T17MOIIgCIK6iBlHEDTKQvrUGASdJWYcQRAEQV3EjCMIgoWbmPF1OTHjCIIgCOoiFEcQBEFQF2GqCoIeQqQlD0oRiiMIFnTChh8sYITiCLqfuDEGwUJFrHEEQRAEdREzjuCDT8xogqAoC82MQ9K2kp6QNE3S0d09niAIgp7KQqE4JPUCzgC2A9YBvihpne4dVRAEQc9koVAcwEhgmu2nbP8XuBTYqZvHFARB0COR7e4eQ7tI2gXY1vaX8/Y+wCa2D6k65kDgwLy5JvBEHV2sCPyz0HBDfsgP+R8c+Qvz2DsjfzXb/ds7aGFZHFeNtnk0nu2zgbM7JVyaZHtEZ84N+SE/5H9w5S/MY2+m/IXFVDUdWLVqeyAwo5vGEgRB0KNZWBTHvcBQSUMkLQ7sAYzv5jEFQRD0SBYKU5Xt2ZIOAa4HegHn2X6kYBedMnGF/JAf8j/w8hfmsTdN/kKxOB4EQRAsOCwspqogCIJgASEURxAEQVAXoTiCIAi6iLxW227bgk4ojiDoBJJ27UhbA/KXlLRmKXk15G/YLNldQU5D1BX9LCKpb0GR/1Oj7YBSwiW9KumVFq+nJV0uaXCxfnrS4rik02gROFiN7UML9fMz4ATgP8B1wDDgcNsXFZLfC7je9tYl5LWQvVFb+23fV6ifPzP/3+J1YBLwG9tvNyh/aeA/tt+X9FFgLeBa2+82IrdK/n22N2qvrZOydwR+Dixue4ik4cBxtsc0Kruqj5uBAcDlwKWFvRSR9AfgPNI1f7+k7Cz/OdL/1mXATS54I5N0CfA14D1gMtAP+KXt/21A5u6kMIItgJurdvUBFrW9ZacHPG8/xwEvApeQAqf3APoD04AvF+unhymO/drab3tsoX6m2B4u6XPAzsARwM22h5WQn/sYD+xju2hu8HxDaQ3b3qpQP78i/aB/n5t2B/4BLAn0tb1Pg/InA5sDywF3kxTSW7b3alDudsBoYDfSTatCX2Ad2yMbkZ/7mAxsBdxie8Pc9qDtDRqV3aKfD5O+x+6k8V9m+4RCsrcG9gc2JSmn820/XkJ2lr8ksCPpxrgRcDVJAd5eQHbl/3cv4GPAd4DJjVx/SUOA1YGfAtXZvd8A7i/4QHO37U1rtUl6oNQ9aKGI4yhFKcXQARbL76OB39t+RaqVNaUh3gYeknQD8GalsdFZU6knkg6woe1PVW3/WdJttj8lqcTTr2y/JekA4DTbP5N0fwG5M0hKaAzpabTCG6QHhBLMtv16E34z82D7H8Cp+WHhKOBHpJlyCdl/Bf4qqR/wReAGSc8DvwUuavRGafs/wDhgnKTlgF8Bt5LivBplMUmLkR76Trf9rqSGnrBtP51nSa/YvrHAGFtF0udtX1n5zNyUTcVmfj1KcVSQ1J/0FLEO0LvSXuppmnQTfJxkqjoo99eQ6aUGf8mvpiFpPea/RhcUEt9f0iDbz+W+BpESsgH8t4B8Sfo4sBdzbcgN/95tPwA8IOmPwJu238ud9QKWaFR+5mFJewK9JA0FDgXuLCQbAElrk2YauwAvkzJOH1m4jxWAvYF9gPuBi4FPAvuRTDaNyv806TtsR8ousVujMjO/AZ4BHgBuk7Qa8K9Ghdp+T9J/JfW13bC8VtgbOE3SOSRT8ERgH0lLAYeX6qRHmaoqSJpAMjN8i2TL3A+YZfs7BftYDvhX/rEsDfTJT3gLBZKOIf1zrwNcQ/rnvN32LoXkjwbOAv5OeiIaAhwE3AJ8xfb/NSj/U6S/7x22T5b0EdI6U6l1rLuBrW3/O28vA0yw/YkCspcCvg+Myk3XAyc0uu7Too+7SWbCy20Xz/sm6UrSutKFJDPVzKp9DSfek/Q0MIU06xhv+812TmkISYvanl1Azu9J5rsJzGsp+GajsruSnqo4Jtv+WLXdWNKttj9dSP5SwDeBQbYPzE+Na9q+uoDscbZ3k/QQNRb6S9nBs/xhJPvrMEkrAefY3rGE/NzHEqSbi4DHS90Y89P/Sba/XUJeK31MsT28vbYFkXx9Lmh0vaedPkbbvqZF2xK23ykguxfwfdvHNSqrFfkrAT8BVra9nVLRuI/bPreA7JoeVCVkZ/krkjy3BlM1w7Z9YGvndIYeaaoCKvbVmZK2J9mtBxaU/zuS/bvy9DmdtEDYsOIADsvvOxSQ1RYVj6TZ2R3xJeAjhfv4GHN/4BtIKmIKy7O8jzUqpx3elLRRxcss9/efEoLzutWutl/L28uRFn63KSE/X58VJC3uVBitGZxAmqlWcxdpIbsh8vi3BJqiOIDzSf/D38/bT5IsFA3f3G2fK2lRYI3cNK3ETKaKq0jOILeTvMKaQk9VHCfkRbsjgdNIHiWlFjYBVre9u6QvQlrIU6GVzsqU3/azJeS1wSRJy5IWMycD/ybZS4sg6UKSl8kU5v7ADZRaQ7k/e55dzrwmgSsLyT8cuFxSxcwzgGRvL8GKFaUBYPtVSR8qJLvCs8Ad+RpVX59fNiI0e2qtAiypFCtS+d33BZZqRHYL7pR0OumGXj3+Eu7iK9oeJ+m7WeZsSUVuwpI2J5nvXiBdmw9L2sf2HSXkA0vbLrpWVYseqTiqTEavA83wIvpvdhc0gKTVgYan6NVI2pSk9NYGFid5k7xpu0iwku2D8sezJF1HcpF9sITszAiS+2qzbKXLkxZ9qx0eDBRRHLbvlbQWqdpkxdRWxKUSeL+F48BqtBF/1Elm5NcipFiCUmwDfIk0g69WQm8A3yvYT2U2Xz3rMPP+vTvLm3lhv/L/uynpXlGCU4DRth/NstcmKZJSxZaulTTK9oRC8mrSU9c4xgKHtTAF/MJ2rajOzsj/LPAD0sLyBGAz4Eu2bykhP/cxieTDfjnpR7cvsIbt77d5Yvty17L9uFoJBCz0RIeky4FDqxdNFyYk9SYt5n+SdIP5G3BWiXUaSduS0mHfmps+BRxo+/pGZdfoa+lmLCxL+oLtP5SW2xXk3/5pwHrAw6R4o11KPDipRjxOrbYG5L9KClh8i+SdKFL81fIl5M/pp4cqjvsrgVVttTXYxwok7wkBd9suWle44pnSYoH/zka9eiSdnRf0awUCupTLcpY/nGT+mjMbc6HoaKVo8TOBlWyvJ2kDYIzLBbiNIz1FV7IBfBFYznaRtCN5kbPy+7mrCb+fj5Ns9svYHiRpGPDVqplmZ+W26R3UqCmsqp+mLWBn+Ysydzb5RKnZpKTzSb/3C3PTXsBSttsMTq5Dfs04lorbeCl6quJ4ANjC9qt5e3ngVtvrF5L/OVIahNfz9rK5vz+VkJ9l3gZsDZxDirieSZrVNBwZKmkR0j9hKbtrrT5qerDZvrVWeyfk3wp8m5S+pBJ9/bDt9QrJny8Kt1ZbnTK7ZLaX+7qHFMMxvuT1yW7crWL7x43Ir+rnWvICdvb6W5TkAdjp/2GlYLlWKbE+lmeqh5JmqgJuIwWoNppiZ6jtqfkBaT4Km5l75hoH8AvS4toVeXtX4MSC8o+x/cfKhu3X8j9UMcVBCqpaBDiEtLC/KvCFEoKzN9XPgY+XkNdKH0UURBssZXtiC5+Ekt4r90va1PbdAJI2ARpVtN8EDiT9PltSyn4/V6D9fIvr0/BTaSnF0AGasYDdlqt5kfUx229LOoUUvPs+MLWQV9XRpEDXM2p1SzJ3FqNHKg7bF+Q1gq1IWv/zlcWqQtTKOlz0Wtt+Vqn++mDSD/qJwq6VEyR9Abiy5AK2pNttf1LSG8y74FuxxZbKRPrP7JRQWeDchTQrK8UmwL5KaSQABgGPKcfXdMZmXfG1d9ekfXle0icA59/RocBjpYTnJ+sDgHWZN/NAkXVEmrCAbXv/EgNri6r1q+dIv/mBkr7S6GK27QPy++aNj7J9epSpSjnUP5um5sP2K4X6OQ94jaT9DXyDZP/+Ugn5uY/tmT/y+qu2ry0k/w1gadJT6H8of2NvKkqR4meTvG9eBZ4G9rb9TCH5q7W1vxF36WxKvRQYZ/vvnZXTTh8rkvI7bU36204gOYy8XEj+5cDjwJ4kz6e9gMdsH9bmiR2X37QF7Cx/e+ZXeg3HjSilIhpj+8m8/VHgKttrNyo7y7uPlBFgXCO/wXb76WGK42rbOyilK6j1tFskwE0pxcgPSf+UkP4pTyzpvZJ/gDvYnpa3Vwf+YnutUn00m7yQtxLzRrg+1/oZnepjaWAR228Ulrs6MN32O5K2ADYgRWO/1vaZHZK9GikmZHeSOeMy0o2g6LVpJhVnk4rzhlLSwOtLOVfkPpq1gH0WKeZkS9Ia4i7AxMpTfYOyb/O8yT1rtjUgf3Xm/nbeIv12Lrf9Qgn5c/rpSYoDQMmou2qz/gnVBekucj/z/Njy97q14A9QpKfEIbaPl7QqMMB2kSBASd8AjiHVDqhk7eyUiacV+UuQ1nwGM69iKhJtLGkKyQ16MCmX1HhSWpnRJeRX9TOU9BCyl+1ixYvU/JoxE22PzE4cB5EcOCYWfDjbFbjO9huSfkCKSD+hhANBlbKrvC9DMtmOavfk9mX/mhTjMo708LorqVbGbQC2xzfaR1Vfa5NiZ75ou6ipvMetcdi2UmbTpqSkcNekuwB4RNI1zPsDvLfiGVLAA+TXpBv6VsDxpMjxM4CNG5Rb4TDSjbaIaaQGV5Fs3pMpHHyZeT8vyH4e+D/bp6lM2nYAlKq1VWplvEdKe16SUbaPyh6A00m/n5uZ617cKGcrxUf9kKRUlyGlbS/FD21fLumTpKDDn5PcrzcpILuSOuYtSSsDr5BMwSXoQ/pdVtLHvEGade9K+j9uWHFIGsjc386izE2dUowepzgyd0va2Pa9TZLf7HQXkGyvLwIVt9ZZpGjpHSnjAbKJ7Y0qN0OntBeLNyizmucpF41bi4G2t22i/HeVUsrsy1xvnMXaOL7DZFfZxUgPBbvafqqE3BY0tWaM7XPyx1spn+MM5nqAbQ+cafsqSccWkn11dqH/GXNrrpzTxvEdxg0WKGsPSXeQlNPlpEJvTzajn56qOLYEvibpGdKNvbLGUarCWlPTXUCXeIC8m81uFa+V/hQoBKO5AWJPAbdI+gvzBgAWCRAjuVuvb/uhQvJasj8pJf+JTkV6hlDgaV0phuaPtk9qVFY7NKVmjKS9bV+kVgIBC/59X5D0G9I64snZNFnLm7HDSNoYeN728Xl7GeAh0iL/KQ2Ot9LH4Cyr4up+B3BkKacNkoPMw4VktUqPW+OA1j1imumFUBo1PzJ6L9JUdyNgLGmB8Ae2L29QblcFiD1KykD6NEkxlX44aBolF0vb6ad4zRhJX7X9m9b+zgX/vksB2wIPOQW+DQDWb8StNXskbZ1nX58iebZ9g5ThYG0XqEUj6S6St9/FuWlP0s2+SMyUpD4k82Dl93Mrae2nrHNIT1QcANk2OtT27/LT1jK2ny4keyDJVXAz0hP77SRXx+kl5Oc+mhoZneWtBXyGdNO90XYxP/8W/SxCuv7FqqI1++FA0mbAscBqpJl7Mc88ST8kzQRaZn4t4i5eo7+zXbheQ1ci6UDbZxeQMyfyX9IZpOJux+btIrVWJN1je5P22hqQfzkpDXylTPY+FFJ61TQ0tVtYyU9D3wG+m5sWo9yiIKRUCOOBlUkppv+c20qyVA0Pp2KR0UqxLi+RfMIvAV7MLpWl5F8iqW9+0n0UeEJSMU80289mJfEfkvKuvEpxLin76ydJDgMjKOc48D/AwSRPm8n5NamQ7FqUysw6B0lj8zpBZXs5pfimZvC1QnJ6ZRdfSA9MN1XtK2XWv0nStyQNlLRKNun9Of8vlIiRGmr7+7afzK8fMrf2RzF66hrH54ANgfsAbM/IU7xS9LddrSjOl1Ss3m+m2ZHR95HSmLxKeppellT46iVSadfJbZ3cAdZxCsbci1Tw5zukG+T/NigXAEljSKk7ViYpwNVIkdHrlpAPvO5CwZYtsV3Kg6ejvNQEmRt4/poixZKItqDUqv7vgVsl/ZP0wPE3AElrUM6RY+/83jIQ8quk/+VBDcp/W9LHbd8FcyLqi5UcrtBTFcd/s1tu5aa7dGH5/5S0N+mHCClzamm304NJttK1JL1AsuWXLAV6HWmR9noASaNINuVxJFfdRqfWi+UZzM7A6bbfrfw9CnE8KbvsX3Mg2pakv0Mpbpb0vySHh+rF/RJxBE0rPVyjr6Wb5H22iKTlPG8i0Wbdb4qUM7Z9oqQbSUW5JniuHX8R0lpHiT5WLSGnDQ4CLszOAiIFAe5bupOeqjjGZY+MZSV9hWQa+G1B+f8DnE7ynjBwZ24rQl4TGGF7azUpMjrLn2MCsD1B0k9sfzP/KBvlN8AzwAPAbXlNotgaB/Cu7ZclLSJpEds3Szq5oPyK4qw285RKRNjM0sMAKOWpOocUX1EsrXoV1YlETYorKJZIVE1Kq+6ctLJFWzGXVkl7ttLvJSXk5weXdbOiVrPipHry4vhngVEkrXy97Ru6eUh10WzPG0kTgBtJniWQPKw+S5p13Gu74drRNfpc1IXqL0v6K2k2cxKwAskcs7EbrFfSFWhurZU5NWLUYMr2Gn00Ja16iz7WYW4i0RtdMJGompBWvSuQdGbVZm/S9Zlsu82U7nXI70syhw1m3owJbdZJqZeeOuMgK4qiykLSBOe0BJK+a/unJeW34AZJ36J5njd7klKC/In0j397butFenrsFO35+TNvudFG2Ilkpz6cZMLrx7xlRjtFF8UpNL30MDQnrXoLlieVM/6dpP6ShpTyXKSJdcGbie2vV29nl+jzC3ZxDWl98iEKxF21Ro9SHJqbyls0J6V3/6rPuwLNVBwV09fBVW2mUJSuU8W51uy60xoQvVR+L+mMMB+238zmr6G2x+Z1gxK5nirrYc0c/zGkNaZVJV1MLj1cuI9mp1U/hmTGW5M0M6h4Lm5WqItm1gXvSt4APlpQ3lK2Dy0oryY9SnHYburNirLunm131GTPmxzbchTzp5Zu1Ia/en5/tNFgwrbIa1cHkp56Vye5RZ9FcrPsNLZ/k9+bVrDI9g1KwWiV0rGHuXDpWJIL669I12U6KYPzwW2eUR/N9lw8kuTyvrpSmo3+JNPbAo1SnrzqRfd1KVvg7RJJ+5PWw6qdNkquH/bMNY78dPJIZUFZKbXAurbvaVDuayTfewGb589zcKF62jX6LR7Aldc4LgO+RbrJ7EcKiPpOg3IfIkWj39OMdZKqfqYAI3M/FRv+Q82wgUu6r+R3UQounJJnTXuTrtevSgUvdgWamx33PqecZ0uTaqcXi9xXk9KqNxNJ1Q8us4FnXS7dCJK+BpxMmslUbu623aib7zz0qBlHFWeS/hkrvFWjrTPsVPX55w3KqofiAVzACrbPlXSYU5nXW5Wi1RvlOuCfwNKSqp+CSheKesf2fys2/HyTadZTUrnsgIkzgWHZ0+nbwHnABcxNaNkwanJadZrsuahU7Ooy4DI3qdhVk7gTeDuHA6wOrCNpeimnENLvZajtZsTmzKFHRo6TZlpzbiK236eAErV9a8sX8EbV52bRjB9J5eltpqTtc/DWwEaF2v627X6kolN9q159CioNSIrue8CS2YPuclIEfzP4S2F5s/PvcyfgVNu/ovyayqhsvtiBZKr6KOmmUwTbPweuAP6QZf/I9mml5ANjSE/s4yTdqxSNXfSpukn8jfSbHEDKI/V10oNBKR6lrFt7TXqq4nhK0qGSFsuvw0jZWptBkXTMLVEqZANAJYCruq0AJ0jqR7Ilf4v0PY4oJdz2nNmZpB1Kya3iaFKq+YdIUbnXAD8o2YGkDytFqN8j6cMFRb+RvYX2Bv6ilKW4WLqXzHxp1QvLh3Tt/0Yy2RbNUuyUUuZntj9G8vbbgBQEu6CziO23SEXGTre9I2nspfgvqazDGZJ+WXkVlA/03DWODwGnknyoTYpXOLwZ07tqX/zCcuezq5e2tXcVC+O4JX2ZVJjoJpKp6tPAcbYbfnrMSmhPUrzM3/KT9Ba2L2hUdlUfJ5HiXP5DWgtaFrja5ZLtNe36VPUxmHmLXV1m+xel5DeDvPb2FZJjwoG2Hy659iapZnnbRgMj5+unJyqOrkTSzraLeU1I2o70lLgbycZboS8p/9PIQv30J/3ABzNvIFGxCPiqvoop17z43uqPutTirKQngE9UInOza+idttcsIb+qnx3chFQjWXZ1WvWlgL5uMK16leymXh/NLXZ1OUlhNMtiUBSl1DffBu5wSnHyEeBbLhexX93XBrYfLC0Xeu7i+BxKPu1KqiXnuUq7C+QxAl4gZUodw9zqZJC8KIqZkkilV/8G/JXygWEt+WpBWc0we9ViOumaV3iDVNWwNMdRMNVIBUn7Vn2u3lVqVtPs67Of7ccLyusSbN9MKtGLpBWzwiuuNDLn07jDT016vOKgrEdMZZrcm+Tp9ECWvwFwDykFd6OcYvszkobZHtv+4Z1mqUZdb9tCuTZ61fZAUgDXQ42YDCsuq0oV+WbafjtvL0mq7VyKF0hrG1eRZjg7ARMrEeWFIsihvMdWheoU8L1J8S33UU5xNOX6VCL3gdGSRrfcX/C6dwUTaNKNPdOs304oDgp6xNjeEkDSpST75UN5ez3SAnMJBkj6NLCdpAtp8eMoNKuBVHd5tO1rCslryQGk8pk35+0tgLuBj0o6zvaFDcq/nLlJAiHNmi6nXM2Mv+dXhavye0PeT0oJLDe1fWduKjkbm4PtebICZEeIRq95NU25PrQdub+w2d2bdmPPFKkGWosevcahlBCs2n5fxLNENaqF1WrrpOxdSDfdTzJ/cR+78cjuSj9vkP5J3yG55haNs5D0Z+DLtl/M2yuR4he+DNzmBpPttfI3KJooMMvsQ7ou/y4o8y4XKiVaR5+LAQ/aXrsJsptR4XEz23e017YgI+kbhV2U58mX11Zbo/TIGYekr5Jsx/8hJQKr5K4qkucJeEzSOaTcPCa5VRbJA2T7CuAKST+0fXwJma300+z0LIMrSiPzEvBRp3rPJSKAZ0kaY3s8gKSdSIGHRcizyAtJKU1QKv6zr+1HCoifIOkLwJVu0pNdVtwV2b2AtUm1VkrJv4SUceA90lpcP0m/tF2kUBepNHNLM0+ttgUOpUWl/sAfJK0MKSVLgzIXJ5kcV8oPM5XZTF8aLw41f389ccYhaSopd3/p/D8V+b1JgT2VtOe3AWdW7O0F+1kOGMq8uaRua/2MBUe+pF+TftCVfFVfIC2ofpvkFrplg/JXBy4mVQAUaWF2HxeKMpZ0Jymld2WhcwvgJy6Qtr1qtvce6eGmdFQ92dxZoZL6YnpB+VNsD1eq8PgxcoXHRr3aJH2cZII8nFTvpkJf4HOlZ5SlkXQQ6aH1ZdLft/K3XadBuUeQin99CHiRuYrjX8Bvbf9fI/Ln66+HKo7rgM/nQJxm9bEkqYLbE02S/2VS+cmBwBRSQry7Cpqqmi1fJGWxGcxJ2/6H0k/YSnnI5MKFrmqZvZphCmsm2TxYWfOZWDKOSdIjwHBSvfrTbd9a4vpkhbcFaTZzVtWuN4A/257aiPxmI2ka6aF1VhNk9wK+bfuk0rLnw3aPe5Gydk4hVaE7tfIqKH8M8ATwdN4eTiqYU/I7PESaCUzJ22uR/NkXCvld/Pe+ugky/wj8kBTnMpgUlf6nQrJFMm/+MG+vCowsPP7dgGeBsSRPqqeBXQrKP5TkWXVN/j6rAX8rKH+17v5ddXLctwC9mij/7q74Hj11xjGR9IQ7T7ETF3JvlTSZFJV+i+dmZn3QZTOD3mt74xyJuontd0otwHeR/M+Tsnh+iHRjKW6OqeqrePR+NuP9mOSkIFLeoR8719huUPaZpN/lVrbXzn1NsF3KI6ySJPCzzrOMHPD5VzdpxpRnmL1crsLjDcCutl/L28sBl9repoT80kiq1MjYgGT+bZn2/NRC/RwPTLJ9VbsHN0CPXBwnJZErWkqxhvzXWwRWlWa6pGVJufxvkPQq0NACWxfL/xmwo+1ixYPa4P7SArOCOBRA0gDbMwuK38QpFfn9lb7y4mdJFvG8pqmXaVLuOklX296BtJZSihUrSgPmXKMPFZRfmkqRt5n5VfwBKXMIyRHhHeZdH1u+ZCc9VXHcLOlAUrbUaq1fKtHbw0pF6XtJGkq6wdzZzjl1Yftz+eOxkm4mlUa9bmGRD7zYRUoD4Eg1Mf0CKRaopDfPu9leXalu15/yZUCvk3Q98Pu8vTvJrNQMVmmCzPclDbL9HIBStccF1nxi+4dd1NWKXdFJTzVV1cqiadtF3HGV8v58HxhF0vjXA0CKCpUAABCESURBVMe7kFdV9ot/0A3GOnSgn+VI9vXqWJciAYaSfgV8mDSjqVbeVxaSfwtprWlR0nrWLODWZsw0S5vCsifS7iRlNJZU2e4HLlwxMbv8VpwTbrP9x5Lyq/o5z4VznEnaFjibZCKE5MF4oO3rS/ZTmuyYs0cLE9tFtrcv2Mdo5np03mK75ANf6qMnKo4PAkq1qL9beeJqgvzjSXWun2Lu065dzqvqdzWaXeoGU7mZZ++wVW0fU3qdqaqvg2z/urDMtUhpQATc2IWzs4aQdKNTSpyT3cSUNbmvFZlbXvcuN8m9viStBKaWTPJ5Iulh4JLctAcpuWTRkgI9ylSlFvmRWlLwaXcE8D3mzyxb8qY1AHgkL/S/WdVHqfK0uwGr2/5vIXnzYHv/ZsitYlGlYjm7kWZ/xZH0SVK1tV9nc9IytkvVhJhK8sFfNPc1qORDQhOdEyopccYopd5pSkqcvNi+LfAR28dJGiRppO2JJeQ3kfckDXSOmVH54lM7Ahvafi/LP4+UgywURwPsmN8/RAoiuilvb0lykyuiOEiBZ9+mhddWCSStQUrW9+MWuz5Ncn8sxcOkGg1Fa5RIOsr2zySdRg2btO1Da5zWGX5MMhHebvtepfTVxXz8JR1DSmS5JvA7Uorvi0hPe43K/gZwDCmQa06QGGUL/jTLOeFHpCJaA4GWCQdN8jYswa/JnmekgLo3SNUGi3meNYkfAXdIqr73fL1wH32BindfUzJA9CjFUXnKlXQ1qXbFzLw9ADijYFeznFNdNIH/A77XcqFX0pukm02pgi0/JVUSe5h51yAandFUblQt82wVIy8sr1o9w3NKX/2Fgt18jhQPdF+WP0Mp1UMJDgPWdK5l0SSa4pzgLkqJQ9d4nhXH9l8kjSQl+BTwHZctIPcz4D5JN2b5W5CUVVF6lOKoYnAL98kXSU+OpThGKVfVjZRf+B1cyzvI9iSlimilGEsyZRSdNdn+c35vWkp4p8JEY5g3JUVp/mvbkiqeT0u3d0IdPE9KMV+cKnPtJEmX0STnBNvH579B9SJtyboiXeF51iy2JJmBT5S0qqSP2Z7c7lkdwPZF2QuyUsnxR7ZLWiKAnqs4bqlyRTRpAenGgvL3J0VaL0bVwjJlTGG929i3ZAH5Ff6/vXOPkauu4vjnSwEpgYJgeSiIBAQpxBKwDRJilIeKSKRWrEUBpaIoaqFRDIEEwQcYjCbyFlBBkadWkYpURd5CeYlCLaI8IgaIKIog8ijHP85vunens2Onc+7dnd3zSTaZuXfn/O7uzNzz+/3O95zzZFRSUidKHOg4PKO4jjjQLZJOx7skVmNAUWXnL5N0DrChpMOBw4Bz+zGo0qsCFyRcJ2kRw2/qEb0m9q88/g+u/FsxBEHbtZJOxlvSXlQOzZdXrz02wj5e7WEhsEkJCL8Pz+Qf05TP5Fq4Q/0y/tk8m9gttl1x0YDhla3DHceEVVVJmsXQbOgpYFMzOzLIdlgP4Q62LwauNbNz247PA95uZnOCxvk6ftO6kuE3r6jg5v10iANZacQUYP/XHQ6HqcLKGPtQkVyb2S/6tHdCl9NmZif1Y79JJP0O2NnMXi7PJwF3RwpEBlF5ptJxtKqkUmCNsxI7nAZcUg69H1hmbf1X+mWirjjAa/O8Gf/HPoQH1qK4VdI0M1saaLPFUcDCovVvLW/fBKyN77tH0ZIH7lY5FhncrDMOhPVZXXcVx/gF0JezaLN3IoCkA9tzNiQdGDVOOwpsn9zGhkArqXaDSMOSvmdmBwPLOhwby7xY8rBaW2wbE7vFtiewk5UVQVFVhSe+TijHIWk7fFtqLl5i4VJ81RV9k9kDOLQkGj7PkNSx79mWeQ+L3eVN71sJgIvM7NouL1udceq+8dYZB2pVfv0K8Goz21fSNLwqaYh4oEY5K8CxDJWb73Ysijpq47TEFb8u9t+C/w1R7Fh9UlY0uwbaD0XSmuZ1us7AJ6lTJZ2IT1zbFZL98Edc0dbq7745rpAMZUJtVUl6GbgRmGdmfyrHHrSgjPHKOFt1Oh61DdM0Gqo1FGnz+3gc6D6GJxhGJQBejctkjzOz6ZLWxLdKQrYQ5eWxQ+WskvYF3oXfTC6tnJqCqwBnRo3VNu6XIhPESo7FFnhtqhm447jNzB4PsH0sniM1GY/RtJzeC8C3AmMooVRXdZJ2BPbGr/2XZhZ2Yy8y35l4G2bwIPktlDifmXXNZVvlcSaY45iFrzh2x+suXQKcZ2ZbB47RSDmQJonMbK3YrC0OVOy3qvtW95Ijq/vebGZ952y02ZyOl+D/Kt4v2vA8jidwVVLflXcrYx0NXG6BzZva7N9pZrWtACSdPFadRCfq+A6NMM5e3c6bWYgIaEJtVZnX4llYpJMHAEfjrRbPAhaa2eKAMV6WdE90pu8oE15dlnrjQADPlv3j1l7vbsRKXOuQsy4FPojHqw7DZ6Rb4iunSCkr+CrmGkn/wCdQV9jwVr79cqukGWZ2e6DNFZjZsTXLfaOZWlHNrUSQYg58dfHfIhXfBk8zWGxB5exbTKgVRyckbQQcCMyJUtyU5eIMoFoOxMzsPRH2m6A41+cqqpg1gHUsqGuipD8A2+DChNA4ULG/C96Deid8j3cq3qgoJFCoGmptSfoGsB6wwErHQklTgK/h78X81bXdZcw34gUVZwOPmtneQXaX4jeth/HvQPT72y73nYv3oRiTqxBJjwFnMUI8qSWMCBjnDtyZbgDcjk/6njKzQyLsrxhnojuOOtDwfs7Cg+VzzWzHEV4y5pB0K7C3mT1Tnq+Hz1z67qld7NUeBypxje3x9+B+M3sxynYdSHoA2M7avpQl8LvMzF5fw5ib4ROnDwDrB97Ya31/m5D7RlKjcq3jOJI+hddOOyVyi7bFhNqqagrz/so7AwcxJPc9u/urxhzrtJwGgJk9Iy8X3zdl9bKojjiQRi5kuZ2kvlVbqrfWlrU7jXJwuUqGehSSPoGvNKYCVwCHR2wbSloH7we+LZ6jc370NkmF2uS+NVBrV7cKa0iagd97PlaOTYoeJB1HIA3KfZvgWUm7WEn4k7Qr3lGsb2qOA+3f5VxEZnSdtbaWSjrEzC6sHpT0ISr5CkFsBRxlZr8NtnsBnq18I7AvnowWvsVG/XLfaLoGrQNZgMt7F5nZvfLinjdGD5JbVYE0JfdtgjJruYShdrGb43GgkJo6I8SBIsvCDxySXoM7tufw5E7D/0eTgVkWUHNI0hQze7rE9lbC+uyCWVXLla3CJdFbNHXKfZNVIx1HIE3IfZtE0loMxQiWRcYI2uJAKzCz6zsd78Fu1w5/UeqVsrr8LCv3XOlbYCFpTzzBTcB9URLKYvsqM3t3SU41hm+hWL+TnPa9/Lr29uuW+w4q8rYLC1j5c/n2kV6zWuOk44inIvedi5cAuIAguW9TlHjGAmArMztc3jt9+zEueazWe9oen422yprsj7dH/WjQOPfgcas78VwLAKJWZHVSZuxb1iEXl7ScoRWkGJ6oZxaTWY+kM4Dv1iX3HVQk/RZvrdD+ubwtdJx0HPVSh9y3CUqOwp3AIWa2k6TJeHvOqAS63XC57A543sIk4NnAG8tiYHZF1ro+nvD2ziD7Az3jHQfXX6vcd1BpSr2VwfGaKXvG55SfQWIbM5sjaS6AmT1XZqpRnI5v612OF2k8BIiUm74WL0PR4gV8+d4XldjATyV9Ei/tXU0A7CtG0CC1Jug1wL6jfQFjlJ9I+hgrfy6fjhwkHUcyEi+UVUYr83obKh/ECMzsT5ImmfdH/o6kWwLNfw9YImkh/jfMAi7s/pJVohW0bjnRz1XOGTAoQoi3AUdIepgBmrE3LPcdRFpbsdXeJIZPpMLIraqkI/JeE8fjcsrFeC/tD5vZdUH2b8ALvZ0HPA48VuyH9CUoY+yKJ1+CxzfqKJ0ykDSRgFkHZQu1Kvd9pI6M+qQ76TiSESm1nnbDZ6O3mtmTgba3wov3rY3XDNsAOLMlYw4aYxKwKcPVJSEBYUlHAheZ2T/L81fi1QHOjLBfF4M+Y29C7jvoyBtcTaPSLdTMfhA6RjqOpBOS5lmld0W5CR/fb02dmpL+Oo3zaeAE3DktJ75W0kplHNRQBdR+GPQZe1Ny30FF0vF4V8o3ANcA7wBusqBy6i0yxpGMxF6SZgPzgI3xCq195VgUfgy0+hL80MxmB9jsxHxcPvz3muyvIUmtEiHFsa5d01iRTKvM2M/HEzAHiemSWoFeAZPL81C57wAzBy/Nf5eZHSxpc2oQ5qTjSDpiZgdJmoNvZ/wH34a5OcB0VZlVZyD5L8SWUW/nGuAySWfjwccj8KTPsc6KJE4zeylWKFc/ZhZed2mc8VypbfZSkaA/Tg3fs3QcSUdKwt98vM3lDsDBZSum37LqNsLjaB4ErpO0iOGyxKi+B58HPg58AneGi/FA/1gnZ+zjm7slbQh8G6+n9jRwV/QgGeNIOiJpGXCkmf2q5G8sAA6zPkvDVzKLq1nFEJ9ZfEKn4/3GaJJkrFK+p5uZ2WPl+bbAlFah0tCx0nEknWgVw2s79noze2C0rmksUVZkJ7OyemVQ8jiScUhTFQHWqHuAZLCQdAx4pqmkA9tOf2QULmm1kDRV0qmSfibp2tZP4BDfwTu6vYQn012IJx0myWiyRN79slbScSTtfKDyuL2/QUidp4a4CO9hsTXen+BhvJVmFJNL1VqZ2SNm9gW8oGWSNE7JaQFPeF0i6X5Jd0m6W1L4VlUGx5N2NMLjTs/HMhub2fmS5pdS7ddLipATt/ivvJPhA/I2nX8FNgm0nyS9sASXuR/QxGDpOJJ2uqmeBikg1pKdPiZpP7wh1RaB9o8C1gU+A3wRX20cGmg/SXpBAGb250YGy+B4UuX/qJ7WMbO1RuvaekHSu/Hs6C3x8u1TgBPN7MquL0ySAUTSo8CIUvNAGTqQK46kjfGSYFVpOPUvPHgdgqSujmcit75NRpVJwHo0tJ2cK45kXCHpNLpsqZnZZ/q0/zc8K/1i4Dbavqj9tr5NktWh6ZpdueJIxht3VB6fiBc6jGQzYB+8LfBBwCLgYjO7L3icJOmFRoUrueJIxi11V6uV9ArcgZwKnGRmp9U1VpJ0Q9JGTXafzBVHMp6pZVZUHMZ+uNN4HfBN4Ed1jJUkq0LTLYvTcSRJD0i6ANgJuBpXad07ypeUJI2TW1XJuELSvxlaaaxLcBFFSS/jcmUYvqLJ6rLJhCEdR5IkSdITWasqSZIk6Yl0HEmSJElPpONIkpqQ9EyXc2+VdNVI55NkLJOOI0mSJOmJdBxJUiNyTpV0r6TfS5pTOT1F0kJJSyWdXcq0J8mYJ/M4kqRe3gvsDEwHXgXcLumGcm4m3nr2EeDn5XevGI2LTJJeyBlOktTLHngtq+Vm9gRwPTCjnFtiZg+a2XK8aOIeo3WRSdIL6TiSpF66FZ8b5EZZyQQmHUeS1MsNwBxJkyRNBd6Ct/kEmClp6xLbmAPcNFoXmSS9kI4jSWpA0prA88BC4HfAPcC1wDFm9nj5td8ApwD3Ag+V302SMU+WHEmSGpA0HTjXzGaO9rUkSTS54kiSYCQdgQe7jx/ta0mSOsgVR5IkSdITueJIkiRJeiIdR5IkSdIT6TiSJEmSnkjHkSRJkvREOo4kSZKkJ9JxJEmSJD3xP2FBVGV7JxDOAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#determining if a variable is good predictor of income through plots\n",
    "%matplotlib inline\n",
    "pd.crosstab(income_c['occupation'],income_c['income']).plot(kind='bar')\n",
    "plt.title('Income against Job Title')\n",
    "plt.xlabel('Job')\n",
    "plt.ylabel('High Salary')\n",
    "\n",
    "pd.crosstab(income_c['education'],income_c['income']).plot(kind='bar')\n",
    "plt.title('Income against Education')\n",
    "plt.xlabel('Education')\n",
    "plt.ylabel('High Salary')\n",
    "\n",
    "table=pd.crosstab(income_c['sex'],income_c['income'])\n",
    "table.div(table.sum(1).astype(float), axis=0).plot(kind='bar',stacked=True)\n",
    "plt.title('Income against Sex')\n",
    "plt.xlabel('Sex')\n",
    "plt.ylabel('High Salary Proportion')\n",
    "\n",
    "table1=pd.crosstab(income_c['race'],income_c['income'])\n",
    "table1.div(table1.sum(1).astype(float), axis=0).plot(kind='bar',stacked=True)\n",
    "plt.title('Income against Race')\n",
    "plt.xlabel('Race')\n",
    "plt.ylabel('High Salary Proportion')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 766,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<=50K    24698\n",
      ">50K      7839\n",
      "Name: income, dtype: int64\n",
      "<=50K    75.91\n",
      ">50K     24.09\n",
      "Name: income, dtype: float64%\n"
     ]
    }
   ],
   "source": [
    "#checking for class imbalance\n",
    "print(income_c['income'].value_counts())\n",
    "print(str(round(income_c['income'].value_counts()*100/income_c.shape[0],2)) + '%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 767,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '# of records')"
      ]
     },
     "execution_count": 767,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#determining if a variable is good predictor of income through plots\n",
    "%matplotlib inline\n",
    "pd.crosstab(income_c['native-country'],income_fact['income']).plot(kind='bar')\n",
    "plt.title('Income against Native Country')\n",
    "plt.xlabel('Native Country')\n",
    "plt.ylabel('# of records')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 768,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['United-States', 'Rest'], dtype=object)"
      ]
     },
     "execution_count": 768,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#since 92% of the records have US as native, adding rest countries into one category called 'Rest'\n",
    "income_c.loc[income_c['native-country']!='United-States', 'native-country'] = 'Rest'\n",
    "income_c['native-country'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 769,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Adm-clerical', 'Exec-managerial', 'Handlers-cleaners',\n",
       "       'Prof-specialty', 'Other-service', 'Sales', 'Craft-repair',\n",
       "       'Transport-moving', 'Farming-fishing', 'Machine-op-inspct',\n",
       "       'Tech-support'], dtype=object)"
      ]
     },
     "execution_count": 769,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Private house service and Protective service was below 1%, hence merged with other services\n",
    "income_c.loc[income_c['occupation']=='Priv-house-serv', 'occupation'] = 'Other-service'\n",
    "income_c.loc[income_c['occupation']=='Protective-serv', 'occupation'] = 'Other-service'\n",
    "income_c.loc[income_c['occupation']=='Armed-Forces', 'occupation'] = 'Other-service'\n",
    "income_c['occupation'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 770,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prof-specialty       18.38\n",
      "Other-service        12.59\n",
      "Craft-repair         12.58\n",
      "Exec-managerial      12.49\n",
      "Adm-clerical         11.58\n",
      "Sales                11.22\n",
      "Machine-op-inspct     6.15\n",
      "Transport-moving      4.91\n",
      "Handlers-cleaners     4.21\n",
      "Farming-fishing       3.05\n",
      "Tech-support          2.85\n",
      "Name: occupation, dtype: float64%\n"
     ]
    }
   ],
   "source": [
    "print(str(round(income_c['occupation'].value_counts()*100/income_c.shape[0],2)) + '%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 771,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "White                 85.43\n",
      "Black                  9.60\n",
      "Asian-Pac-Islander     3.19\n",
      "Amer-Indian-Eskimo     0.96\n",
      "Other                  0.83\n",
      "Name: race, dtype: float64%\n"
     ]
    }
   ],
   "source": [
    "print(str(round(income_c['race'].value_counts()*100/income_c.shape[0],2)) + '%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 772,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "White                 85.43\n",
      "Black                  9.60\n",
      "Asian-Pac-Islander     3.19\n",
      "Other                  1.79\n",
      "Name: race, dtype: float64%\n"
     ]
    }
   ],
   "source": [
    "#American-Indian-Eskimo and Other were below 1%, hence merged with other\n",
    "income_c.loc[income_c['race']=='Amer-Indian-Eskimo', 'race'] = 'Other'\n",
    "income_c['race'].unique()\n",
    "print(str(round(income_c['race'].value_counts()*100/income_c.shape[0],2)) + '%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 773,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '# of records')"
      ]
     },
     "execution_count": 773,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#determining if a variable is good predictor of income through plots\n",
    "%matplotlib inline\n",
    "pd.crosstab(income_c['marital-status'],income_fact['income']).plot(kind='bar')\n",
    "plt.title('Income against Marital Status')\n",
    "plt.xlabel('Marital Status')\n",
    "plt.ylabel('# of records')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 774,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Never-married', 'Married', 'Divorced', 'Separated', 'Widowed'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 774,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#all married-regardless of absentee spouse(1.28%), Af-spouse(0.07%), Civ-spouse can be merged into one -married\n",
    "#Grouping within married-spouse would hardly help explain variation due to sparse presence of 2 of the 3 groups\n",
    "income_c.loc[income_c['marital-status']=='Married-AF-spouse', 'marital-status'] = 'Married'\n",
    "income_c.loc[income_c['marital-status']=='Married-spouse-absent', 'marital-status'] = 'Married'\n",
    "income_c.loc[income_c['marital-status']=='Married-civ-spouse', 'marital-status'] = 'Married'\n",
    "income_c['marital-status'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 775,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Married          47.36\n",
      "Never-married    32.78\n",
      "Divorced         13.65\n",
      "Separated         3.15\n",
      "Widowed           3.05\n",
      "Name: marital-status, dtype: float64%\n",
      "Private             75.33\n",
      "Self-emp-not-inc     7.81\n",
      "Local-gov            6.43\n",
      "State-gov            3.99\n",
      "Self-emp-inc         3.43\n",
      "Federal-gov          2.95\n",
      "Without-pay          0.04\n",
      "Never-worked         0.02\n",
      "Name: workclass, dtype: float64%\n"
     ]
    }
   ],
   "source": [
    "print(str(round(income_c['marital-status'].value_counts()*100/income_c.shape[0],2)) + '%')\n",
    "print(str(round(income_c['workclass'].value_counts()*100/income_c.shape[0],2)) + '%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 776,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '# of records')"
      ]
     },
     "execution_count": 776,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "pd.crosstab(income_c['relationship'],income_fact['income']).plot(kind='bar')\n",
    "plt.title('Income against Work Class')\n",
    "plt.xlabel('Work-class')\n",
    "plt.ylabel('# of records')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 777,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['State-gov', 'Self-emp-not-inc', 'Private', 'Federal-gov',\n",
       "       'Local-gov', 'Self-emp-inc', 'No Pay/Job'], dtype=object)"
      ]
     },
     "execution_count": 777,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#although distribution of private is dominating but behavior across other types is varying\n",
    "#merged never worked and no pay dut to similar income status\n",
    "\n",
    "income_c.loc[income_c['workclass']=='Without-pay', 'workclass'] = 'No Pay/Job'\n",
    "income_c.loc[income_c['workclass']=='Never-worked', 'workclass'] = 'No Pay/Job'\n",
    "income_c['workclass'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 808,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Education')"
      ]
     },
     "execution_count": 808,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "pd.crosstab(income_c['income'],income_c['education']).plot(kind='bar')\n",
    "plt.title('income and Education')\n",
    "plt.xlabel('income')\n",
    "plt.ylabel('Education')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating Dummy Variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 778,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Creting an array of column names which are to be converted to factors\n",
    "factors = ['workclass', 'marital-status', 'occupation', 'relationship', 'race', 'sex', 'native-country', 'income']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 779,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['age', 'fnlwgt', 'education', 'capital-gain', 'capital-loss',\n",
       "       'hours-per-week', 'income', 'workclass_Federal-gov',\n",
       "       'workclass_Local-gov', 'workclass_No Pay/Job', 'workclass_Private',\n",
       "       'workclass_Self-emp-inc', 'workclass_Self-emp-not-inc',\n",
       "       'workclass_State-gov', 'marital-status_Divorced',\n",
       "       'marital-status_Married', 'marital-status_Never-married',\n",
       "       'marital-status_Separated', 'marital-status_Widowed',\n",
       "       'occupation_Adm-clerical', 'occupation_Craft-repair',\n",
       "       'occupation_Exec-managerial', 'occupation_Farming-fishing',\n",
       "       'occupation_Handlers-cleaners', 'occupation_Machine-op-inspct',\n",
       "       'occupation_Other-service', 'occupation_Prof-specialty',\n",
       "       'occupation_Sales', 'occupation_Tech-support',\n",
       "       'occupation_Transport-moving', 'relationship_Husband',\n",
       "       'relationship_Not-in-family', 'relationship_Other-relative',\n",
       "       'relationship_Own-child', 'relationship_Unmarried', 'relationship_Wife',\n",
       "       'race_Asian-Pac-Islander', 'race_Black', 'race_Other', 'race_White',\n",
       "       'sex_Female', 'sex_Male', 'native-country_Rest',\n",
       "       'native-country_United-States'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 779,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for var in factors:\n",
    "    if var != 'income':\n",
    "        cat_list='var'+'_'+var\n",
    "        cat_list = pd.get_dummies(income_c[var], prefix=var)\n",
    "        income_c = pd.concat([income_c,cat_list], axis = 1)\n",
    "        income_c = income_c.drop(var, 1)\n",
    "income_c.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 803,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Education')"
      ]
     },
     "execution_count": 803,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "pd.crosstab(income_c['race_White'],income_c['education']).plot(kind='bar')\n",
    "plt.title('Race and Education')\n",
    "plt.xlabel('Race')\n",
    "plt.ylabel('Education')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 780,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Dropping one dummy variable from each categorical var\n",
    "income_c=income_c.drop(['race_Other','sex_Male','native-country_Rest','relationship_Unmarried',\n",
    "                      'occupation_Other-service','workclass_No Pay/Job','marital-status_Never-married'],axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test-Training Set Split and Fitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 781,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='lbfgs',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 781,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "#70% records being sent into training set \n",
    "income_c['income'] = income_c['income'].factorize()[0]\n",
    "pts = np.random.rand(len(income_c)) < 0.7\n",
    "\n",
    "income_train = income_c[pts]\n",
    "income_test = income_c[~pts]\n",
    "\n",
    "#Prepping the training set\n",
    "X_train = income_train.copy()\n",
    "X_train = X_train.drop('income', 1)\n",
    "Y_train = income_train['income']\n",
    "\n",
    "#Prepping the test set\n",
    "X_test = income_test.copy()\n",
    "X_test = X_test.drop('income', 1)\n",
    "Y_test = income_test['income']\n",
    "\n",
    "#FULL MODEL - all variables w/o scaling and feature selection\n",
    "#fitting the model\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "logreg = LogisticRegression(solver='lbfgs')\n",
    "logreg.fit(X_train, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 648,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[7120  278]\n",
      " [1724  624]]\n",
      "accuracy score 0.7945823927765236\n",
      "f1_score 0.384\n"
     ]
    },
    {
     "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": [
    "#generating confusion matrix\n",
    "#generating confusion matrix\n",
    "X_test = income_test.copy()\n",
    "X_test = X_test.drop('income', 1)\n",
    "Y_test = income_test['income']\n",
    "Y_pred = logreg.predict(X_test)\n",
    "\n",
    "#Confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(Y_test, Y_pred)\n",
    "sns.heatmap(cm, annot=True)\n",
    "print(cm)\n",
    "\n",
    "#Testing set accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "print('accuracy score',accuracy_score(Y_test, Y_pred))\n",
    "\n",
    "#Precision=True positive / (True positive + False positive)\n",
    "#Recall=True positive / (True positive + False negative)\n",
    "#F1 is the harmonic mean of precision and recall\n",
    "from sklearn.metrics import f1_score\n",
    "print('f1_score',f1_score(Y_test, Y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Scaling (Standardization)\n",
    "\n",
    "Capital gains and capital losses have wide range\n",
    "<br>Similarly, scale for age and other quantities are also varying \n",
    "<br>We are getting convergence warning on using feature selection methods like RFE and RandomForest Classifier\n",
    "<br>As we are not using our model for interpretation but prediction of log odds,\n",
    "<br>To resolve, all of the above we are standardizing our dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 782,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\shali\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\shali\\Anaconda3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n",
      "C:\\Users\\shali\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:7: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  import sys\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X=StandardScaler()\n",
    "#Normalizing the training set \n",
    "X_train_sc=pd.DataFrame(sc_X.fit_transform(X_train))\n",
    "X_train_sc.columns=X_train.columns\n",
    "#Normalizing the test set\n",
    "X_test_sc=pd.DataFrame(sc_X.transform(X_test))\n",
    "X_test_sc.columns=X_test.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Selection\n",
    "<h4> 1. RFE(Recursive Feature Elimination)</h4>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 783,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num Features: 18\n",
      "Selected Features: [ True False  True  True  True  True  True  True  True  True  True  True\n",
      " False  True False False False False  True  True False False False False\n",
      " False False  True False False  True  True False False False  True False]\n",
      "Feature Ranking: [ 1 17  1  1  1  1  1  1  1  1  1  1  2  1 12  9  8  6  1  1 10 19  4  5\n",
      "  3 16  1 13 18  1  1 15 14 11  1  7]\n"
     ]
    }
   ],
   "source": [
    "target=Y_train\n",
    "# feature extraction\n",
    "from sklearn.feature_selection import RFE\n",
    "model = LogisticRegression(solver='lbfgs')\n",
    "rfe = RFE(model)\n",
    "fit = rfe.fit(X_train_sc, target)\n",
    "print(\"Num Features: %d\" % fit.n_features_)\n",
    "print(\"Selected Features: %s\" % fit.support_)\n",
    "print(\"Feature Ranking: %s\" % fit.ranking_)\n",
    "\n",
    "#Creating a training set with important features\n",
    "X_train_rfe=X_train_sc[X_train_sc.columns[fit.support_]]\n",
    "#Creating a test set with important features\n",
    "X_test_rfe=X_test_sc[X_test_sc.columns[fit.support_]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 784,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[6920  505]\n",
      " [1032 1341]]\n",
      "accuracy score 0.8431312512757706\n",
      "f1_score 0.6356956624792605\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "logreg = LogisticRegression(solver='lbfgs')\n",
    "logreg.fit(X_train_rfe, Y_train)\n",
    "Y_pred_rfe = logreg.predict(X_test_rfe)\n",
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(Y_test, Y_pred_rfe)\n",
    "sns.heatmap(cm,annot=True,cbar=False)\n",
    "plt.ylabel('True Label')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.title('Confusion Matrix')\n",
    "print(cm)\n",
    "\n",
    "#Testing set accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "print('accuracy score',accuracy_score(Y_test, Y_pred_rfe))\n",
    "\n",
    "#Precision=True positive / (True positive + False positive)\n",
    "#Recall=True positive / (True positive + False negative)\n",
    "#F1 is the harmonic mean of precision and recall\n",
    "from sklearn.metrics import f1_score\n",
    "print('f1_score',f1_score(Y_test, Y_pred_rfe))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h4>2. RFECV(Recursive Feature Elimination with Cross-Fold Validation)</h4>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 785,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\shali\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\shali\\Anaconda3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n",
      "C:\\Users\\shali\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:7: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  import sys\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X=StandardScaler()\n",
    "#Normalizing the training set \n",
    "X_train_sc=pd.DataFrame(sc_X.fit_transform(X_train))\n",
    "X_train_sc.columns=X_train.columns\n",
    "#Normalizing the test set\n",
    "X_test_sc=pd.DataFrame(sc_X.transform(X_test))\n",
    "X_test_sc.columns=X_test.columns\n",
    "\n",
    "logreg = LogisticRegression(solver='lbfgs')\n",
    "\n",
    "from sklearn.feature_selection import RFECV\n",
    "rfecv = RFECV(estimator=logreg,  cv=4, scoring='accuracy')\n",
    "rfecv=rfecv.fit(X_train_sc, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 786,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of features: 32\n",
      "Number of features selected: 32\n",
      "Test Accuracy 0.844049806082874\n"
     ]
    }
   ],
   "source": [
    "print('Optimal number of features: {}'.format(rfecv.n_features_))\n",
    "cols = X_train_sc.columns[rfecv.support_]\n",
    "acc = accuracy_score(Y_test, rfecv.estimator_.predict(X_test_sc[cols]))\n",
    "print('Number of features selected: {}'.format(rfecv.n_features_))\n",
    "print('Test Accuracy {}'.format(acc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 787,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(16, 9))\n",
    "plt.title('Recursive Feature Elimination with Cross-Validation', fontsize=18, fontweight='bold', pad=20)\n",
    "plt.xlabel('Number of features selected', fontsize=14, labelpad=20)\n",
    "plt.ylabel('% Correct Classification', fontsize=14, labelpad=20)\n",
    "plt.plot(range(1, len(rfecv.grid_scores_) + 1), rfecv.grid_scores_, color='#303F9F', linewidth=3)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear Regression (Q5,6)\n",
    "<h4>Q5.</h4>Use the income dataset to regress work hours on the rest of the data. Compare the R^2 with R^2 of another model that you fit with only the significant variables. Explore some interaction terms and explain why they were of interest."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 849,
   "metadata": {},
   "outputs": [],
   "source": [
    "#70% records being sent into training set \n",
    "income_c['hours-per-week'] = income_c['hours-per-week'].factorize()[0]\n",
    "pts = np.random.rand(len(income_c)) < 0.7\n",
    "\n",
    "Wh_train = income_c[pts]\n",
    "Wh_test = income_c[~pts]\n",
    "\n",
    "#Prepping the training set\n",
    "X_train = Wh_train.copy()\n",
    "X_train = Wh_train.drop('hours-per-week', 1)\n",
    "Y_train = Wh_train['hours-per-week']\n",
    "\n",
    "#Prepping the test set\n",
    "X_test = Wh_test.copy()\n",
    "X_test = Wh_test.drop('hours-per-week', 1)\n",
    "Y_test = Wh_test['hours-per-week']"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://www.statsmodels.org/dev/generated/statsmodels.regression.linear_model.OLS.html> ![image.png](attachment:image.png)\n",
    "#No constant is added by the model unless you are using formulas in ols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 850,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.040\n",
      "Model:                            OLS   Adj. R-squared:                  0.038\n",
      "Method:                 Least Squares   F-statistic:                     26.05\n",
      "Date:                Thu, 16 Jan 2020   Prob (F-statistic):          1.00e-169\n",
      "Time:                        22:23:02   Log-Likelihood:                -89622.\n",
      "No. Observations:               22795   AIC:                         1.793e+05\n",
      "Df Residuals:                   22758   BIC:                         1.796e+05\n",
      "Df Model:                          36                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "================================================================================================\n",
      "                                   coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------------------------\n",
      "const                            7.3360      3.207      2.287      0.022       1.050      13.622\n",
      "age                              0.0580      0.008      7.505      0.000       0.043       0.073\n",
      "fnlwgt                       -2.225e-06   7.97e-07     -2.791      0.005   -3.79e-06   -6.62e-07\n",
      "education                       -0.1556      0.051     -3.078      0.002      -0.255      -0.057\n",
      "capital-gain                  1.433e-05   1.17e-05      1.226      0.220   -8.59e-06    3.73e-05\n",
      "capital-loss                     0.0002      0.000      0.926      0.354      -0.000       0.001\n",
      "income                          -0.5776      0.238     -2.426      0.015      -1.044      -0.111\n",
      "workclass_Federal-gov           -3.3097      3.131     -1.057      0.290      -9.447       2.827\n",
      "workclass_Local-gov             -1.7776      3.112     -0.571      0.568      -7.877       4.322\n",
      "workclass_Private               -0.7937      3.095     -0.256      0.798      -6.861       5.273\n",
      "workclass_Self-emp-inc           1.0549      3.127      0.337      0.736      -5.075       7.185\n",
      "workclass_Self-emp-not-inc       1.8862      3.106      0.607      0.544      -4.202       7.974\n",
      "workclass_State-gov             -0.3674      3.122     -0.118      0.906      -6.486       5.751\n",
      "marital-status_Divorced         -1.0254      0.306     -3.355      0.001      -1.625      -0.426\n",
      "marital-status_Married          -0.7969      0.606     -1.315      0.188      -1.985       0.391\n",
      "marital-status_Separated        -0.5281      0.507     -1.041      0.298      -1.522       0.466\n",
      "marital-status_Widowed           2.7349      0.557      4.911      0.000       1.643       3.826\n",
      "occupation_Adm-clerical         -3.3526      0.341     -9.839      0.000      -4.020      -2.685\n",
      "occupation_Craft-repair         -4.2546      0.344    -12.382      0.000      -4.928      -3.581\n",
      "occupation_Exec-managerial      -2.5660      0.350     -7.325      0.000      -3.253      -1.879\n",
      "occupation_Farming-fishing       0.6475      0.547      1.184      0.236      -0.424       1.719\n",
      "occupation_Handlers-cleaners    -2.3940      0.469     -5.100      0.000      -3.314      -1.474\n",
      "occupation_Machine-op-inspct    -4.9795      0.411    -12.128      0.000      -5.784      -4.175\n",
      "occupation_Prof-specialty       -0.5923      0.325     -1.825      0.068      -1.228       0.044\n",
      "occupation_Sales                -1.1198      0.349     -3.210      0.001      -1.804      -0.436\n",
      "occupation_Tech-support         -2.4561      0.552     -4.452      0.000      -3.537      -1.375\n",
      "occupation_Transport-moving     -2.2791      0.446     -5.108      0.000      -3.154      -1.405\n",
      "relationship_Husband             0.6441      0.669      0.963      0.336      -0.667       1.955\n",
      "relationship_Not-in-family       0.4071      0.319      1.277      0.202      -0.218       1.032\n",
      "relationship_Other-relative      0.2791      0.562      0.496      0.620      -0.823       1.381\n",
      "relationship_Own-child           2.4141      0.375      6.431      0.000       1.678       3.150\n",
      "relationship_Wife                1.4667      0.742      1.977      0.048       0.013       2.920\n",
      "race_Asian-Pac-Islander          0.6675      0.778      0.857      0.391      -0.858       2.193\n",
      "race_Black                      -0.4851      0.673     -0.721      0.471      -1.805       0.835\n",
      "race_White                       0.7804      0.623      1.252      0.210      -0.441       2.002\n",
      "sex_Female                       1.1362      0.244      4.655      0.000       0.658       1.615\n",
      "native-country_United-States     0.3476      0.322      1.078      0.281      -0.284       0.980\n",
      "==============================================================================\n",
      "Omnibus:                    12468.445   Durbin-Watson:                   1.987\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):            88858.229\n",
      "Skew:                           2.612   Prob(JB):                         0.00\n",
      "Kurtosis:                      11.140   Cond. No.                     2.17e+07\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 2.17e+07. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n"
     ]
    }
   ],
   "source": [
    "#FULL MODEL - all variables w/o scaling and feature selection\n",
    "#fitting the model\n",
    "#No constant is added by the model unless you are using formulas\n",
    "import statsmodels.api as sm\n",
    "\n",
    "#Adding constant for beta-zero(intercept)\n",
    "X_train=sm.add_constant(X_train)\n",
    "model=sm.OLS(list(Y_train),X_train.astype(float))\n",
    "results=model.fit()\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 852,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.033\n",
      "Model:                            OLS   Adj. R-squared:                  0.032\n",
      "Method:                 Least Squares   F-statistic:                     48.08\n",
      "Date:                Thu, 16 Jan 2020   Prob (F-statistic):          1.04e-150\n",
      "Time:                        22:23:34   Log-Likelihood:                -89704.\n",
      "No. Observations:               22795   AIC:                         1.794e+05\n",
      "Df Residuals:                   22778   BIC:                         1.796e+05\n",
      "Df Model:                          16                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "================================================================================================\n",
      "                                   coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------------------------\n",
      "const                            7.1751      0.497     14.447      0.000       6.202       8.149\n",
      "age                              0.0540      0.007      7.521      0.000       0.040       0.068\n",
      "fnlwgt                       -2.673e-06   7.88e-07     -3.390      0.001   -4.22e-06   -1.13e-06\n",
      "education                       -0.2347      0.045     -5.237      0.000      -0.323      -0.147\n",
      "workclass_Private               -0.6825      0.200     -3.419      0.001      -1.074      -0.291\n",
      "marital-status_Divorced         -0.8553      0.253     -3.384      0.001      -1.351      -0.360\n",
      "marital-status_Widowed           2.9210      0.517      5.650      0.000       1.908       3.934\n",
      "occupation_Adm-clerical         -3.2479      0.276    -11.772      0.000      -3.789      -2.707\n",
      "occupation_Craft-repair         -3.8050      0.277    -13.727      0.000      -4.348      -3.262\n",
      "occupation_Exec-managerial      -2.1215      0.265     -8.020      0.000      -2.640      -1.603\n",
      "occupation_Handlers-cleaners    -2.0971      0.429     -4.884      0.000      -2.939      -1.255\n",
      "occupation_Machine-op-inspct    -4.6374      0.363    -12.768      0.000      -5.349      -3.925\n",
      "occupation_Tech-support         -2.2488      0.509     -4.420      0.000      -3.246      -1.252\n",
      "occupation_Transport-moving     -1.9623      0.400     -4.901      0.000      -2.747      -1.177\n",
      "relationship_Own-child           2.3497      0.256      9.166      0.000       1.847       2.852\n",
      "race_White                       1.1648      0.235      4.961      0.000       0.705       1.625\n",
      "sex_Female                       1.2369      0.202      6.122      0.000       0.841       1.633\n",
      "==============================================================================\n",
      "Omnibus:                    12513.429   Durbin-Watson:                   1.987\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):            89740.068\n",
      "Skew:                           2.622   Prob(JB):                         0.00\n",
      "Kurtosis:                      11.185   Cond. No.                     1.55e+06\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 1.55e+06. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n"
     ]
    }
   ],
   "source": [
    "#Dropping most insignificant columns\n",
    "#Race_white is the prominent feature, other races are insignificant\n",
    "#Only widowed seems to be significant enough in marital-status\n",
    "#Capital-loss and capital -gain gave insignificant results as they have high p-values\n",
    "#Income is also highly insignificant\n",
    "#occupation-Farming-Fishing and Prof-Specialty have very high p-value (insignificant)\n",
    "\n",
    "X_train_new=X_train\n",
    "X_train_new=X_train_new.drop(['capital-gain','capital-loss','income',\n",
    "                              'marital-status_Married','marital-status_Separated',\n",
    "                             'occupation_Prof-specialty','occupation_Farming-fishing','occupation_Sales',\n",
    "                              'race_Asian-Pac-Islander','race_Black',\n",
    "                             'relationship_Wife','relationship_Husband','relationship_Other-relative',\n",
    "                             'relationship_Not-in-family','native-country_United-States',\n",
    "                             'workclass_Federal-gov','workclass_Local-gov',\n",
    "                              'workclass_Self-emp-inc','workclass_Self-emp-not-inc','workclass_State-gov']\n",
    "                             ,1)\n",
    "import statsmodels.formula.api as sm\n",
    "model2=sm.OLS(list(Y_train),X_train_new.astype(float))\n",
    "results=model2.fit()\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 853,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.033\n",
      "Model:                            OLS   Adj. R-squared:                  0.032\n",
      "Method:                 Least Squares   F-statistic:                     45.25\n",
      "Date:                Thu, 16 Jan 2020   Prob (F-statistic):          6.99e-150\n",
      "Time:                        22:24:42   Log-Likelihood:                -89704.\n",
      "No. Observations:               22795   AIC:                         1.794e+05\n",
      "Df Residuals:                   22777   BIC:                         1.796e+05\n",
      "Df Model:                          17                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "================================================================================================\n",
      "                                   coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------------------------\n",
      "const                            7.0654      0.600     11.777      0.000       5.890       8.241\n",
      "age                              0.0540      0.007      7.527      0.000       0.040       0.068\n",
      "fnlwgt                       -2.671e-06   7.89e-07     -3.387      0.001   -4.22e-06   -1.13e-06\n",
      "education                       -0.2008      0.113     -1.770      0.077      -0.423       0.022\n",
      "workclass_Private               -0.6808      0.200     -3.409      0.001      -1.072      -0.289\n",
      "marital-status_Divorced         -0.8565      0.253     -3.388      0.001      -1.352      -0.361\n",
      "marital-status_Widowed           2.9194      0.517      5.646      0.000       1.906       3.933\n",
      "occupation_Adm-clerical         -3.2510      0.276    -11.776      0.000      -3.792      -2.710\n",
      "occupation_Craft-repair         -3.8085      0.277    -13.729      0.000      -4.352      -3.265\n",
      "occupation_Exec-managerial      -2.1222      0.265     -8.022      0.000      -2.641      -1.604\n",
      "occupation_Handlers-cleaners    -2.0985      0.429     -4.886      0.000      -2.940      -1.257\n",
      "occupation_Machine-op-inspct    -4.6375      0.363    -12.768      0.000      -5.349      -3.926\n",
      "occupation_Tech-support         -2.2523      0.509     -4.426      0.000      -3.250      -1.255\n",
      "occupation_Transport-moving     -1.9646      0.400     -4.906      0.000      -2.750      -1.180\n",
      "relationship_Own-child           2.3493      0.256      9.164      0.000       1.847       2.852\n",
      "race_White                       1.2894      0.449      2.874      0.004       0.410       2.169\n",
      "sex_Female                       1.2378      0.202      6.126      0.000       0.842       1.634\n",
      "edu*race                        -0.0393      0.121     -0.326      0.744      -0.275       0.197\n",
      "==============================================================================\n",
      "Omnibus:                    12512.596   Durbin-Watson:                   1.987\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):            89721.680\n",
      "Skew:                           2.621   Prob(JB):                         0.00\n",
      "Kurtosis:                      11.184   Cond. No.                     1.89e+06\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 1.89e+06. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n"
     ]
    }
   ],
   "source": [
    "#Adding an interaction term for education and race\n",
    "X_train_new['edu*race']=X_train['race_White']*X_train['education']\n",
    "#dropping the previous interaction term -race*sex\n",
    "#dropping the previous interaction term -age*sex\n",
    "\n",
    "import statsmodels.formula.api as sm\n",
    "model2=sm.OLS(list(Y_train),X_train_new.astype(float))\n",
    "results=model2.fit()\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 858,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.033\n",
      "Model:                            OLS   Adj. R-squared:                  0.032\n",
      "Method:                 Least Squares   F-statistic:                     45.81\n",
      "Date:                Thu, 16 Jan 2020   Prob (F-statistic):          7.81e-152\n",
      "Time:                        22:28:45   Log-Likelihood:                -89700.\n",
      "No. Observations:               22795   AIC:                         1.794e+05\n",
      "Df Residuals:                   22777   BIC:                         1.796e+05\n",
      "Df Model:                          17                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "================================================================================================\n",
      "                                   coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------------------------\n",
      "const                            8.4115      0.641     13.114      0.000       7.154       9.669\n",
      "age                              0.0242      0.012      1.990      0.047       0.000       0.048\n",
      "fnlwgt                       -2.723e-06   7.89e-07     -3.453      0.001   -4.27e-06   -1.18e-06\n",
      "education                       -0.6260      0.136     -4.601      0.000      -0.893      -0.359\n",
      "workclass_Private               -0.6734      0.200     -3.374      0.001      -1.065      -0.282\n",
      "marital-status_Divorced         -0.8612      0.253     -3.408      0.001      -1.356      -0.366\n",
      "marital-status_Widowed           3.0053      0.518      5.806      0.000       1.991       4.020\n",
      "occupation_Adm-clerical         -3.2296      0.276    -11.705      0.000      -3.770      -2.689\n",
      "occupation_Craft-repair         -3.8026      0.277    -13.720      0.000      -4.346      -3.259\n",
      "occupation_Exec-managerial      -2.1177      0.264     -8.007      0.000      -2.636      -1.599\n",
      "occupation_Handlers-cleaners    -2.1521      0.430     -5.008      0.000      -2.994      -1.310\n",
      "occupation_Machine-op-inspct    -4.6476      0.363    -12.798      0.000      -5.359      -3.936\n",
      "occupation_Tech-support         -2.1923      0.509     -4.307      0.000      -3.190      -1.195\n",
      "occupation_Transport-moving     -1.9492      0.400     -4.869      0.000      -2.734      -1.164\n",
      "relationship_Own-child           2.2712      0.258      8.817      0.000       1.766       2.776\n",
      "race_White                       1.1531      0.235      4.912      0.000       0.693       1.613\n",
      "sex_Female                       1.2565      0.202      6.217      0.000       0.860       1.653\n",
      "age*edu                          0.0094      0.003      3.046      0.002       0.003       0.016\n",
      "==============================================================================\n",
      "Omnibus:                    12513.639   Durbin-Watson:                   1.987\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):            89769.534\n",
      "Skew:                           2.621   Prob(JB):                         0.00\n",
      "Kurtosis:                      11.187   Cond. No.                     1.81e+06\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 1.81e+06. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n"
     ]
    }
   ],
   "source": [
    "#The interaction term is highly insignificant\n",
    "X_train_new=X_train_new.drop('edu*race',1)\n",
    "#creating an interaction term for age and education\n",
    "X_train_new['age*edu']=X_train['age']*X_train['education']\n",
    "\n",
    "import statsmodels.formula.api as sm\n",
    "model2=sm.OLS(list(Y_train),X_train_new.astype(float))\n",
    "results=model2.fit()\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h4>Conclusion:</h4>\n",
    "As, all the variables except 'age' turned out to be significant. We stopped our process here. We need more features to further improve our model because much of our variation is explained by our intercept ( as can be seen by our high beta-zero coefficient). The existing set of features are significant but not enough."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h4>Q6.</h4>Regress work hours of only people who are working in Sales. Once you fit a reasonable (as per your judgement) model for this data, compare the R^2 with the ones you calculated for Question 5. Explain your findings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 860,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.028\n",
      "Model:                            OLS   Adj. R-squared:                  0.026\n",
      "Method:                 Least Squares   F-statistic:                     10.62\n",
      "Date:                Thu, 16 Jan 2020   Prob (F-statistic):           6.06e-18\n",
      "Time:                        22:49:31   Log-Likelihood:                -13963.\n",
      "No. Observations:                3650   AIC:                         2.795e+04\n",
      "Df Residuals:                    3639   BIC:                         2.802e+04\n",
      "Df Model:                          10                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "================================================================================================\n",
      "                                   coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------------------------\n",
      "const                           12.6212      1.487      8.491      0.000       9.707      15.536\n",
      "age                             -0.0198      0.029     -0.678      0.498      -0.077       0.037\n",
      "fnlwgt                       -1.577e-06   1.78e-06     -0.885      0.376   -5.07e-06    1.92e-06\n",
      "education                       -1.0402      0.310     -3.352      0.001      -1.649      -0.432\n",
      "workclass_Private               -1.0656      0.497     -2.143      0.032      -2.041      -0.091\n",
      "marital-status_Divorced         -1.3831      0.602     -2.299      0.022      -2.563      -0.203\n",
      "marital-status_Widowed           0.2705      1.204      0.225      0.822      -2.089       2.630\n",
      "occupation_Adm-clerical        1.57e-16   5.37e-17      2.921      0.004    5.16e-17    2.62e-16\n",
      "occupation_Craft-repair      -5.739e-16   2.51e-16     -2.284      0.022   -1.07e-15   -8.13e-17\n",
      "occupation_Exec-managerial    4.435e-16   6.04e-17      7.338      0.000    3.25e-16    5.62e-16\n",
      "occupation_Handlers-cleaners   5.39e-17   7.25e-18      7.438      0.000    3.97e-17    6.81e-17\n",
      "occupation_Machine-op-inspct          0          0        nan        nan           0           0\n",
      "occupation_Tech-support               0          0        nan        nan           0           0\n",
      "occupation_Transport-moving           0          0        nan        nan           0           0\n",
      "relationship_Own-child           2.9808      0.554      5.385      0.000       1.896       4.066\n",
      "race_White                      -0.7473      0.589     -1.270      0.204      -1.901       0.407\n",
      "sex_Female                       1.5634      0.439      3.561      0.000       0.703       2.424\n",
      "age*edu                          0.0159      0.007      2.133      0.033       0.001       0.030\n",
      "==============================================================================\n",
      "Omnibus:                     1693.729   Durbin-Watson:                   2.062\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             8467.120\n",
      "Skew:                           2.242   Prob(JB):                         0.00\n",
      "Kurtosis:                       8.963   Cond. No.                     1.07e+39\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The smallest eigenvalue is 1.5e-64. This might indicate that there are\n",
      "strong multicollinearity problems or that the design matrix is singular.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\shali\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  after removing the cwd from sys.path.\n"
     ]
    }
   ],
   "source": [
    "#Filtering for people who are working in Sales\n",
    "#70% records being sent into training set\n",
    "income_s=income_c.loc[income_c['occupation_Sales'] == 1]\n",
    "income_s['hours-per-week'] = income_s['hours-per-week'].factorize()[0]\n",
    "\n",
    "#Prepping the training set\n",
    "X_sales = income_s.copy()\n",
    "X_sales = income_s.drop('hours-per-week', 1)\n",
    "Y_sales = income_s['hours-per-week']\n",
    "\n",
    "\n",
    "#Dropping columns as per our final model\n",
    "X_sales=X_sales.drop(['capital-gain','capital-loss','income',\n",
    "                              'marital-status_Married','marital-status_Separated',\n",
    "                             'occupation_Prof-specialty','occupation_Farming-fishing','occupation_Sales',\n",
    "                              'race_Asian-Pac-Islander','race_Black',\n",
    "                             'relationship_Wife','relationship_Husband','relationship_Other-relative',\n",
    "                             'relationship_Not-in-family','native-country_United-States',\n",
    "                             'workclass_Federal-gov','workclass_Local-gov',\n",
    "                              'workclass_Self-emp-inc','workclass_Self-emp-not-inc','workclass_State-gov'],1)\n",
    "\n",
    "#adding the interaction term for education & age\n",
    "X_sales['age*edu']=X_sales['education']*X_sales['age']\n",
    "\n",
    "#Running our model\n",
    "import statsmodels.api as sm\n",
    "#adding a constant for beta-zero(intercept)\n",
    "X_sales=sm.add_constant(X_sales)\n",
    "model_sales=sm.OLS(list(Y_sales),X_sales.astype(float))\n",
    "results=model_sales.fit()\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 861,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.028\n",
      "Model:                            OLS   Adj. R-squared:                  0.026\n",
      "Method:                 Least Squares   F-statistic:                     10.62\n",
      "Date:                Thu, 16 Jan 2020   Prob (F-statistic):           6.06e-18\n",
      "Time:                        22:51:38   Log-Likelihood:                -13963.\n",
      "No. Observations:                3650   AIC:                         2.795e+04\n",
      "Df Residuals:                    3639   BIC:                         2.802e+04\n",
      "Df Model:                          10                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "===============================================================================================\n",
      "                                  coef    std err          t      P>|t|      [0.025      0.975]\n",
      "-----------------------------------------------------------------------------------------------\n",
      "const                          12.6212      1.487      8.491      0.000       9.707      15.536\n",
      "age                            -0.0198      0.029     -0.678      0.498      -0.077       0.037\n",
      "fnlwgt                      -1.577e-06   1.78e-06     -0.885      0.376   -5.07e-06    1.92e-06\n",
      "education                      -1.0402      0.310     -3.352      0.001      -1.649      -0.432\n",
      "workclass_Private              -1.0656      0.497     -2.143      0.032      -2.041      -0.091\n",
      "marital-status_Divorced        -1.3831      0.602     -2.299      0.022      -2.563      -0.203\n",
      "marital-status_Widowed          0.2705      1.204      0.225      0.822      -2.089       2.630\n",
      "occupation_Transport-moving  6.483e-16   2.48e-16      2.613      0.009    1.62e-16    1.13e-15\n",
      "relationship_Own-child          2.9808      0.554      5.385      0.000       1.896       4.066\n",
      "race_White                     -0.7473      0.589     -1.270      0.204      -1.901       0.407\n",
      "sex_Female                      1.5634      0.439      3.561      0.000       0.703       2.424\n",
      "age*edu                         0.0159      0.007      2.133      0.033       0.001       0.030\n",
      "==============================================================================\n",
      "Omnibus:                     1693.729   Durbin-Watson:                   2.062\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             8467.120\n",
      "Skew:                           2.242   Prob(JB):                         0.00\n",
      "Kurtosis:                       8.963   Cond. No.                     7.87e+22\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The smallest eigenvalue is 2.78e-32. This might indicate that there are\n",
      "strong multicollinearity problems or that the design matrix is singular.\n"
     ]
    }
   ],
   "source": [
    "#Trying another model for sales people\n",
    "\n",
    "#Removing other occupations as we have selected only those working in Sales -they are hardly making any difference\n",
    "#Going with the assumption that government roles doesn't involve sales role(or very few)\n",
    "#having it as separate variables won't add much value to the model\n",
    "X_sales=X_sales.drop(['occupation_Adm-clerical','occupation_Craft-repair',\n",
    "                    'occupation_Exec-managerial','occupation_Handlers-cleaners',\n",
    "                    'occupation_Machine-op-inspct','occupation_Tech-support'],1)\n",
    "import statsmodels.api as sm\n",
    "model_sales=sm.OLS(list(Y_sales),X_sales.astype(float))\n",
    "results=model_sales.fit()\n",
    "print(results.summary())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 862,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.026\n",
      "Model:                            OLS   Adj. R-squared:                  0.024\n",
      "Method:                 Least Squares   F-statistic:                     15.98\n",
      "Date:                Thu, 16 Jan 2020   Prob (F-statistic):           3.24e-18\n",
      "Time:                        22:55:02   Log-Likelihood:                -13968.\n",
      "No. Observations:                3650   AIC:                         2.795e+04\n",
      "Df Residuals:                    3643   BIC:                         2.799e+04\n",
      "Df Model:                           6                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "===============================================================================================\n",
      "                                  coef    std err          t      P>|t|      [0.025      0.975]\n",
      "-----------------------------------------------------------------------------------------------\n",
      "const                          11.1445      0.597     18.664      0.000       9.974      12.315\n",
      "education                      -0.4192      0.103     -4.059      0.000      -0.622      -0.217\n",
      "workclass_Private              -1.3684      0.480     -2.849      0.004      -2.310      -0.427\n",
      "marital-status_Divorced        -1.1991      0.590     -2.031      0.042      -2.357      -0.041\n",
      "marital-status_Widowed          0.6913      1.137      0.608      0.543      -1.538       2.921\n",
      "occupation_Transport-moving  3.792e-16    3.8e-16      0.997      0.319   -3.66e-16    1.12e-15\n",
      "relationship_Own-child          2.7486      0.500      5.497      0.000       1.768       3.729\n",
      "sex_Female                      1.4933      0.431      3.468      0.001       0.649       2.338\n",
      "==============================================================================\n",
      "Omnibus:                     1704.958   Durbin-Watson:                   2.061\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             8597.853\n",
      "Skew:                           2.257   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.014   Cond. No.                     3.01e+19\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The smallest eigenvalue is 6.79e-35. This might indicate that there are\n",
      "strong multicollinearity problems or that the design matrix is singular.\n"
     ]
    }
   ],
   "source": [
    "#Checking the model on further removal of insignificant variables\n",
    "X_sales=X_sales.drop(['race_White','fnlwgt','age','age*edu'],1)#,'fnlwgt','race_White','education'],1)\n",
    "import statsmodels.formula.api as sm\n",
    "model_sales=sm.OLS(list(Y_sales),X_sales.astype(float))\n",
    "results=model_sales.fit()\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On further removing insignificant variable, more and more variables turned out to be insignificant. Our R-squared is very low and unstable to use this model even for interpretation. \n",
    "<br>We need extra information to further explain the variation in work-hours."
   ]
  }
 ],
 "metadata": {
  "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}