ETA on Airline Delay Improving Sampling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Flight ETA Challenge: Arrival time data for Airlines\n",
    "    Author: Rishabh Shrivas"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Task: \n",
    "Build a binary classification model for predicting arrival delays > 15minutes without using departure delay as an input feature."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dataset: \n",
    "US Dept. of Transportation on-time performance data for non-stop domestic flights\n",
    "    - Link: https://transtats.bts.gov/DL_SelectFields.asp?Table_ID=236&DB_Short_Name=On-Time\n",
    "    - Data range: [Geography = All of America], [Year-Month = 2017 Jan, Apr and 2016 Jul, Oct]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Approach: \n",
    "    1) Load data\n",
    "    2) Wrangle data: Check for class imbalances and handle it by sampling from the dataset\n",
    "    3) Wrangle data: Check for sparse columns to remove them\n",
    "    4) Wrangle data: Check for columns with a few missing values to impute with zero, mean or median\n",
    "    5) Wrangle data: Encode categorical values to numerical values\n",
    "    6) Exploratory analyss: Infer distribution of data using histograms\n",
    "    7) Exploratory analyss: Use measures of central tendency and IQR to understand data at various positions\n",
    "    8) Variable interdependence: Infer multicollinearity of data through correlation matrix\n",
    "    9) Feature selection: Remove redundant variables (i.e. columns that refer to the same thing e.g. ID)\n",
    "    10) Feature engineering: Remove highly collinear variables (i.e. correlation > 0.8)\n",
    "    11) Feature engineering: Reduce data dimension to the variables that most explain variation in target variable\n",
    "    12) Modeling building: Random forest is chosen for machine learning modeling\n",
    "    13) Model comparison: Random forest modeling with raw data (PART 2) and PCA transformed data (PART 3)\n",
    "    14) Machine learning process:\n",
    "        (a) Train/test data: Split data for training and testing\n",
    "        (b) Model creation: Create models in a grid search for a combination of parameters\n",
    "        (c) Model selection: Use 3-fold cross validation to evaluate model performance and select the best one\n",
    "        (c) Final Model training: Use parameters of the best model from above to train on entire train set\n",
    "        (d) Final Model evaluation: Use 3-fold cross validation for average error on 3 folds of training data\n",
    "        (e) Final Model performance evaluation: Use confusion matrix, Precision, Recall, ROC to evaluate performance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PART 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Environment set up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\homeuser\\Anaconda3\\lib\\site-packages\\sklearn\\cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import csv\n",
    "import os\n",
    "import tabulate\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "import matplotlib\n",
    "from matplotlib import pyplot as plt\n",
    "import matplotlib.pyplot as plt\n",
    "from datetime import datetime\n",
    "from sklearn.preprocessing import Imputer\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn import linear_model, decomposition, datasets\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn import cross_validation\n",
    "from sklearn.metrics import confusion_matrix, roc_curve\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "%matplotlib inline\n",
    "get_ipython().magic('matplotlib inline')\n",
    "matplotlib.rcParams.update({'font.size': 12})\n",
    "matplotlib.rc('xtick', labelsize=8) \n",
    "matplotlib.rc('ytick', labelsize=8)\n",
    "np.random.seed(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Set Up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "startTime = datetime.now()\n",
    "folderPath = \"C:/Users/homeuser/Documents/MEJ/DS Challenge/StartUpML/Flight Delay/dataFiles\"\n",
    "os.chdir(folderPath)\n",
    "fdata = pd.DataFrame()\n",
    "for filename in os.listdir(os.getcwd()):#Reads all data files from above location\n",
    "    temp = pd.read_csv(filename)\n",
    "    fdata = fdata.append(temp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "folderPath = \"C:/Users/homeuser/Documents/MEJ/DS Challenge/StartUpML/Flight Delay\"\n",
    "os.chdir(folderPath)#Setting working directory\n",
    "#fileName = \"On_Time_On_Time_Performance_2017_1.csv\"\n",
    "#fdata = pd.read_csv(fileName)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Wrangling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Target variable is 'ArrDel15' and there was substantial class imbalances as seen below. In order to avoid the classifier learning one class better than the other, sampling was done to even out the class imbalance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Class imbalance:\n",
      "0.0    1491392\n",
      "1.0     366913\n",
      "Name: ArrDel15, dtype: int64\n",
      "Class imbalance evened out:\n",
      "1.0    366913\n",
      "0.0    366913\n",
      "Name: ArrDel15, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "110"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Handles class imbalance through sampling\n",
    "classDistribution = fdata['ArrDel15'].value_counts()\n",
    "print('Class imbalance:')\n",
    "print(classDistribution)\n",
    "fdata = fdata.sample(frac=1).reset_index(drop=True)\n",
    "zero = fdata[fdata['ArrDel15']==0].tail(classDistribution.min())\n",
    "one = fdata[fdata['ArrDel15']==1]\n",
    "data = zero.append(one)\n",
    "del zero, one\n",
    "data = data.sample(frac=1).reset_index(drop=True)\n",
    "#data.sort(['Year','Month','DayofMonth','DayOfWeek'], ascending=[False,False,False,False])\n",
    "#del fdata\n",
    "print('Class imbalance evened out:')\n",
    "print(data['ArrDel15'].value_counts())\n",
    "len(data.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Year</th>\n",
       "      <th>Quarter</th>\n",
       "      <th>Month</th>\n",
       "      <th>DayofMonth</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>FlightDate</th>\n",
       "      <th>UniqueCarrier</th>\n",
       "      <th>AirlineID</th>\n",
       "      <th>Carrier</th>\n",
       "      <th>TailNum</th>\n",
       "      <th>...</th>\n",
       "      <th>Div4TailNum</th>\n",
       "      <th>Div5Airport</th>\n",
       "      <th>Div5AirportID</th>\n",
       "      <th>Div5AirportSeqID</th>\n",
       "      <th>Div5WheelsOn</th>\n",
       "      <th>Div5TotalGTime</th>\n",
       "      <th>Div5LongestGTime</th>\n",
       "      <th>Div5WheelsOff</th>\n",
       "      <th>Div5TailNum</th>\n",
       "      <th>Unnamed: 109</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>24</td>\n",
       "      <td>1</td>\n",
       "      <td>2017-04-24</td>\n",
       "      <td>F9</td>\n",
       "      <td>20436</td>\n",
       "      <td>F9</td>\n",
       "      <td>N713FR</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>13</td>\n",
       "      <td>3</td>\n",
       "      <td>2016-07-13</td>\n",
       "      <td>UA</td>\n",
       "      <td>19977</td>\n",
       "      <td>UA</td>\n",
       "      <td>N466UA</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 110 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Year  Quarter  Month  DayofMonth  DayOfWeek  FlightDate UniqueCarrier  \\\n",
       "0  2017        2      4          24          1  2017-04-24            F9   \n",
       "1  2016        3      7          13          3  2016-07-13            UA   \n",
       "\n",
       "   AirlineID Carrier TailNum      ...       Div4TailNum  Div5Airport  \\\n",
       "0      20436      F9  N713FR      ...               NaN          NaN   \n",
       "1      19977      UA  N466UA      ...               NaN          NaN   \n",
       "\n",
       "   Div5AirportID  Div5AirportSeqID Div5WheelsOn Div5TotalGTime  \\\n",
       "0            NaN               NaN          NaN            NaN   \n",
       "1            NaN               NaN          NaN            NaN   \n",
       "\n",
       "  Div5LongestGTime  Div5WheelsOff Div5TailNum  Unnamed: 109  \n",
       "0              NaN            NaN         NaN           NaN  \n",
       "1              NaN            NaN         NaN           NaN  \n",
       "\n",
       "[2 rows x 110 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head(2)#Showing top 2 rows as sampled data is ordered by time with older data at the top"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Year</th>\n",
       "      <th>Quarter</th>\n",
       "      <th>Month</th>\n",
       "      <th>DayofMonth</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>FlightDate</th>\n",
       "      <th>UniqueCarrier</th>\n",
       "      <th>AirlineID</th>\n",
       "      <th>Carrier</th>\n",
       "      <th>TailNum</th>\n",
       "      <th>...</th>\n",
       "      <th>Div4TailNum</th>\n",
       "      <th>Div5Airport</th>\n",
       "      <th>Div5AirportID</th>\n",
       "      <th>Div5AirportSeqID</th>\n",
       "      <th>Div5WheelsOn</th>\n",
       "      <th>Div5TotalGTime</th>\n",
       "      <th>Div5LongestGTime</th>\n",
       "      <th>Div5WheelsOff</th>\n",
       "      <th>Div5TailNum</th>\n",
       "      <th>Unnamed: 109</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>733824</th>\n",
       "      <td>2016</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>2016-10-01</td>\n",
       "      <td>NK</td>\n",
       "      <td>20416</td>\n",
       "      <td>NK</td>\n",
       "      <td>N615NK</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>733825</th>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>2017-04-02</td>\n",
       "      <td>NK</td>\n",
       "      <td>20416</td>\n",
       "      <td>NK</td>\n",
       "      <td>N617NK</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 110 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Year  Quarter  Month  DayofMonth  DayOfWeek  FlightDate UniqueCarrier  \\\n",
       "733824  2016        4     10           1          6  2016-10-01            NK   \n",
       "733825  2017        2      4           2          7  2017-04-02            NK   \n",
       "\n",
       "        AirlineID Carrier TailNum      ...       Div4TailNum  Div5Airport  \\\n",
       "733824      20416      NK  N615NK      ...               NaN          NaN   \n",
       "733825      20416      NK  N617NK      ...               NaN          NaN   \n",
       "\n",
       "        Div5AirportID  Div5AirportSeqID Div5WheelsOn Div5TotalGTime  \\\n",
       "733824            NaN               NaN          NaN            NaN   \n",
       "733825            NaN               NaN          NaN            NaN   \n",
       "\n",
       "       Div5LongestGTime  Div5WheelsOff Div5TailNum  Unnamed: 109  \n",
       "733824              NaN            NaN         NaN           NaN  \n",
       "733825              NaN            NaN         NaN           NaN  \n",
       "\n",
       "[2 rows x 110 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.tail(2)#Showing bottom 2 rows as sampled data is ordered by time with latest data at the bottom"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Sparsity: \n",
    "Columns with >80% missing values are worth not pursuing for analysis and cannot be imputed sensibly as well"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "56"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = data.iloc[:,0:56]#Removing sparse columns based on the information given in DOT website\n",
    "print('Dimension reduced to:')\n",
    "len(data.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "56\n",
      "Sparsity per variable:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Year                    0.0\n",
       "Quarter                 0.0\n",
       "Month                   0.0\n",
       "DayofMonth              0.0\n",
       "DayOfWeek               0.0\n",
       "FlightDate              0.0\n",
       "UniqueCarrier           0.0\n",
       "AirlineID               0.0\n",
       "Carrier                 0.0\n",
       "TailNum                 0.0\n",
       "FlightNum               0.0\n",
       "OriginAirportID         0.0\n",
       "OriginAirportSeqID      0.0\n",
       "OriginCityMarketID      0.0\n",
       "Origin                  0.0\n",
       "OriginCityName          0.0\n",
       "OriginState             0.0\n",
       "OriginStateFips         0.0\n",
       "OriginStateName         0.0\n",
       "OriginWac               0.0\n",
       "DestAirportID           0.0\n",
       "DestAirportSeqID        0.0\n",
       "DestCityMarketID        0.0\n",
       "Dest                    0.0\n",
       "DestCityName            0.0\n",
       "DestState               0.0\n",
       "DestStateFips           0.0\n",
       "DestStateName           0.0\n",
       "DestWac                 0.0\n",
       "CRSDepTime              0.0\n",
       "DepTime                 0.0\n",
       "DepDelay                0.0\n",
       "DepDelayMinutes         0.0\n",
       "DepDel15                0.0\n",
       "DepartureDelayGroups    0.0\n",
       "DepTimeBlk              0.0\n",
       "TaxiOut                 0.0\n",
       "WheelsOff               0.0\n",
       "WheelsOn                0.0\n",
       "TaxiIn                  0.0\n",
       "CRSArrTime              0.0\n",
       "ArrTime                 0.0\n",
       "ArrDelay                0.0\n",
       "ArrDelayMinutes         0.0\n",
       "ArrDel15                0.0\n",
       "ArrivalDelayGroups      0.0\n",
       "ArrTimeBlk              0.0\n",
       "Cancelled               0.0\n",
       "CancellationCode        1.0\n",
       "Diverted                0.0\n",
       "CRSElapsedTime          0.0\n",
       "ActualElapsedTime       0.0\n",
       "AirTime                 0.0\n",
       "Flights                 0.0\n",
       "Distance                0.0\n",
       "DistanceGroup           0.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Dimension reduced to:')\n",
    "print(len(data.columns))\n",
    "print('Sparsity per variable:')\n",
    "((len(data.index) - data.count())/len(data.index))#Sparsity per variable retained"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.drop(['DepDelay','DepDel15','ArrDelay','DepDelayMinutes','ArrDelayMinutes'],\n",
    "          axis=1, inplace=True)#To be excluded as per the task\n",
    "data.drop(['CancellationCode'], axis=1, inplace=True)#Removed due to 100% sparsity as seen in above cell\n",
    "data[\"FlightDate\"] = data[\"FlightDate\"].apply(lambda x: int(''.join(x.split(\"-\"))))#Formatting date for convinience"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Datatype of variables:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Year                      int64\n",
       "Quarter                   int64\n",
       "Month                     int64\n",
       "DayofMonth                int64\n",
       "DayOfWeek                 int64\n",
       "FlightDate                int64\n",
       "UniqueCarrier            object\n",
       "AirlineID                 int64\n",
       "Carrier                  object\n",
       "TailNum                  object\n",
       "FlightNum                 int64\n",
       "OriginAirportID           int64\n",
       "OriginAirportSeqID        int64\n",
       "OriginCityMarketID        int64\n",
       "Origin                   object\n",
       "OriginCityName           object\n",
       "OriginState              object\n",
       "OriginStateFips           int64\n",
       "OriginStateName          object\n",
       "OriginWac                 int64\n",
       "DestAirportID             int64\n",
       "DestAirportSeqID          int64\n",
       "DestCityMarketID          int64\n",
       "Dest                     object\n",
       "DestCityName             object\n",
       "DestState                object\n",
       "DestStateFips             int64\n",
       "DestStateName            object\n",
       "DestWac                   int64\n",
       "CRSDepTime                int64\n",
       "DepTime                 float64\n",
       "DepartureDelayGroups    float64\n",
       "DepTimeBlk               object\n",
       "TaxiOut                 float64\n",
       "WheelsOff               float64\n",
       "WheelsOn                float64\n",
       "TaxiIn                  float64\n",
       "CRSArrTime                int64\n",
       "ArrTime                 float64\n",
       "ArrDel15                float64\n",
       "ArrivalDelayGroups      float64\n",
       "ArrTimeBlk               object\n",
       "Cancelled               float64\n",
       "Diverted                float64\n",
       "CRSElapsedTime          float64\n",
       "ActualElapsedTime       float64\n",
       "AirTime                 float64\n",
       "Flights                 float64\n",
       "Distance                float64\n",
       "DistanceGroup             int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Datatype of variables:')\n",
    "data.dtypes#Checking variable data type as numeric type is most easy to deal with"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exploratory Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0312d2080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAGMCAYAAABtZVBoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nThgqJElSJ/4/9i2TLxb5A5MAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a001afd278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAGFCAYAAABKagGgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a001748c88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAGMCAYAAABtZVBoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nyXGlVAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0017abf28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAG7CAYAAAB0Gmd9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzs3Xe8HGXZ//HPl4A0JZTQ1IQQUQiCSoIN6WJBKSoIRkAC\nKMUCxsfy/CwUEbHwPLEgEnwEIkgPqKCo9GIB5YQeEjDAiRiBQ4mUhBKu3x/3nGSz7DnZnZ3ds7Pn\n+3699nV2Z2av675PTmavnbnnHkUEZmZmZs1aYagbYGZmZt3BRYWZmZkVwkWFmZmZFcJFhZmZmRXC\nRYWZmZkVwkWFmZmZFcJFhZmZmRXCRYWZmZkVwkWFmZmZFcJFhZm1jaQdJL0k6eihbksRsr5cXbXs\n2Gz59kPVLrOh4qLCrINlH06Vj0WSHpF0i6SfSXq/pLb/P5Z0TI22DfaYW/H2yB4dQ9IZNdr8oqQ+\nSVdJ+ngD4Tquf2btsuJQN8DMliuAYwEBI4A1gTcC+wOHAH+XtF9E3NvGNl3Dyz843wLsCdwG/Kpq\n3ZPZz5uA8UBfS1uXTwC/Bm7NXr8CGAfsAewkaXxEfGOoGmdWBi4qzEogIo6vXiZpXeDHwD7AFZK2\njoi2fFhHxPXA9VXtORD4EHBrRHxzgPctAua0voW5/SoiflG5QNJWwC3AFyQdHxHPD03TzDqfT3+Y\nlVREPApMAq4FRgNfrd5G0lqSTpR0t6RnJT0p6UpJ76mx7YHZYf9PSPqgpD9JelrS45IulLRJs20e\naEyFpGslLZa0oqSjJd0naaGkeyR9smK7wyXdnvVlnqRjB8n1dkkXSZov6TlJvZJOlbRhI22OiJnA\n48AqwKsa6/Ey7Rkj6a7sFNZ+eeOYdTIfqTArsYgISd8CdiQVGF/oXydpDHAdMAa4AbgcWB3YDfi9\npEMj4ufVIYG9gF2Bi0mnOd6SLdtR0jYtOs3SfyrlPOBtwO+AF4C9gWmSXgDeDBwAXAZcSTotcbSk\nZyLi+5XBJB0MTAMWAb8B5gGvJ50u2l3S2yPin/U0TNIEYG3g/oh4LE/nJL0569PqwK4RcU2eOGad\nzkWFWfndCLwIrCdpo4h4MFv+C9IRjI9FxIX9G0tag1Rs/EjSb7IjHktWk4qO3SLi8or3fA74IXAK\n8LKjHAVR1t43RsRTWd7/Be4BpgJPAFtGxL+zdccB9wFflPQ/EfFStvz1wE+BucAO/dtn63YCrsj6\nsleNNnxY0sbZ81cAGwO7k4qSA3J1StoFuAh4Ctg2Iu7ME8esDHz6w6zksnP8/d+g1wWQ9CZge2BG\nZUGRbf8f4BjS4fxaH6xXVRYUmZ8A/wB2ljS6wOYv0zTgK/0FRdbW+0lF00jgm5UFQkQsAC4FRgGv\nqYjzadIXps9Xbp+95xrSkYvdJa1elV9kRz+yx38D+2bLzwHuarRDkvYHfksqSt7hgsK6nY9UmHUH\nZT/7TyO8M/s5UtIxNbZfL3vP+Brrrq9eEBEvSbqRdDXEVqQPyVa4pcayf2U/e2qseyj7+VqWtukd\n2c8dJb2txnvWI11F8wZgZsXyACZHxFkAkpTFPRA4DtgzGwz7bJ19+TzpapgbgD2yYs6sq7moMCs5\nSSuTzvkD9J/KWCf7+R4GPl0RpHP81R4eYPv+b/0jG21jvSqPUlR4Mfu5YJB1K1Us6+/7FwdLBbyy\nxnIt2SAiSIXKtyRtCnwc+Bzw3UHiVsbZLnt+tQsKGy58+sOs/LYjfUF4OCJ6s2X9H8BHRcSIQR6f\nrBFv/QHybFAVu1P1t2+NQfq9YkTc0EDMm0iFQq0jH7UEaVDoTcCx2fgPs67nosKsxLJD9F8jfYj9\nsmLVX7Of273sTcu3Q408KwDbZi9nVq/vMP19L3Ka7LWyn43sM58kHSW6AfiGpO8U2B6zjuSiwqyk\nJK0HnE8qAh4ETuxfFxG3kD7MPiLpoAHev0U2gVa1nSV9sGrZ54DXkQ7lt2o8RVFOJp0WmZpdCbIM\nSStJ2vblb6tN0lrAwaTCraFLQSPiGeD9pEtgvyxpaiPvNysbj6kwK4GKwZYrsHSa7m1JYwn+Cuwf\nEY9Xve3jwFXA/0k6knQo/knS4MM3ZTHeydJxGP0uBS6RdAnpks2tSB+MfcBniu1Z8SJidjZPxc+B\nuyT9njSL50qkOTu2Ax4BNq96q1j2ktIRpN/V7qQxKzeT5r5otD0LJe0GzACOkrRqRBzeeM/MOp+L\nCrPO1n81R/8MlM+T5jt4EJgOXBQRV9R8Y8RDkiaSjjLsRSoyRpAGXN5Nmqvhjhr5LgZOI51W+QBp\nEqqLgK9GxH11tHd5N9MaaJvB3tfQDboi4peSbgX+C9iJdBriGdKVJBeSjvDUyrFH9uj3FGmejBOB\nk2tM0V3XzcMi4nlJHyZN7vUpSa+IiIMb6ZNZGSgNcDaz4S67d8fpwEHV978wM6uHx1SYmZlZITqy\nqJC0SXYzo9mSbpL0sgl6JL1X0kxJPdnPhyT9fSjaa9ZFtPxNzMxq69QxFdOAUyPiLEl7kc4dL3N9\neET8Efhj/2tJl5IGpZlZfj4fama5ddyYiuwSt3uBtStuEDQfeFdEzB3gPa8mjVIfExF9bWusmZmZ\nLdGJpz9GA/P7C4pML+lSsIEcCPzWBYWZmdnQ6dTTH406GPjsQCslrQO8D3gAWNSmNpmZmXWDVYCx\nwB8i4rHBNuzEomIesKGkFSqOVowhHa14GUk7AitTMb6ihvex7BTGZmZm1pj9gHMG26DjioqIeFRS\nD3AAMF3S3sC8gcZTkI5SnBmDDw55AODss89m/Phad3qubcqUKUyd2rpZdVsdvx053Iehj9+OHO5D\nZ+RwHzojx3Drw6xZs9h///0h+ywdTMcVFZnDgTMlfZV0x8HJANmd/h6KiNOy12sAHwa2XE68RQDj\nx49nwoQJdTdi5MiRDW3fqFbHb0cO92Ho47cjh/vQGTnch87IMYz7sNzhAx1ZVETEHGCbGsuPqXr9\nH+BV7WqXmZmZDawTr/4wMzOzEnJRYWZmZoUYceyxxw51G1ruuOOO2xA47LDDDmPDDTds6L1bbrm8\n4RrNaXX8duRwH4Y+fjtyuA+dkcN96Iwcw6kP8+fP57TTTgM47dhjj50/2LYdN6NmK0iaANxyyy23\ntHxwjZmZWTfp6elh4sSJABMjomewbX36w8zMzArhosLMzMwK4aLCzMzMCuGiwszMzArhosLMzMwK\n4aLCzMzMCuGiwszMzArhosLMzMwK0dQNxSS9A9gJWA84JSLulbQasBkwJyKeLqCNZmZmVgK5jlRI\neoWki4E/AScARwKjs9UvAX8EjiqkhWZmZlYKeU9/HA/sBhwBbAqof0VELAIuBPZsunVmZmZWGnmL\niknATyPiNODxGutnAeNyt8rMzMxKJ29RsR5wxyDrFwOr5YxtZmZmJZS3qJhHGow5kHcB9+WMbWZm\nZiWUt6g4BzhM0jsrlgWApE8B+wC/aLJtZmZmViJ5Lyk9AXgHcD1p/EQAUyWtDbwW+B0wtZAWmpmZ\nWSnkOlIREc8D7wcOAuYC9wArA7cDk4HdI2JxQW00MzOzEsg9+VVEBHB29jAzM7NhLu/kV3Ml7THI\n+t0kzc3fLDMzMyubvAM1xwKvHGT9K4GNcsY2MzOzEmrmhmIxyLq3Ak82EdvMzMxKpu4xFZKOYun9\nPAL4gaQTamw6EliTdNmpmZmZDRONDNR8BLgrez4WeCh7VArgGeAW4JRmG2dmZmblUXdRERHnAucC\nSLoG+FZEXNWqhpmZmVm55LqkNCJ2KrohZmZmVm6556kAkLQS6R4gI6kx6DMirm8mvpmZmZVHrqJC\n0grAicCnGfxupCPyxDczs3Lr7e2lr6+v7u1HjRrFmDFjWtgia4e8Ryq+CnwJmAbcCJwFfIV0Gemn\nSQM2v1xEA83MrFx6e3vZdNPxLFr0bN3vWWWV1Zg9e5YLi5LLW1RMBi6IiCMkrZMtuyUirpY0HfgL\nsDNwZQFtNDOzgrXySEJfX19WUJwNjK/jHbNYtGh/+vr6GioqfDSk8+QtKl4LfC97/lz2cxVINxuT\ndDbwBdIRDTMz6yDtO5IwHpjQcPvq4aMhnSlvUfEY2TTdEfG0pP8A46q2WauZhpmZWWu060hCK3VD\nH7pR3qJiJmkq7n7XAJ+XNJN0FciRwG15GyVpE2A6MIo0TmNyRMyqsd1o4CfAG4AXgZ9GxE/y5jUz\nG15adyShfbqhD90j770/TgNWlrRy9vprpKm5rweuA9YA/quJdk0DTo2ITUmnWaYPsN0lwJkRsVlE\nbAFc0EROMzMza0Leya9+A/ym4vXdkl4H7AgsBv4cEY/niS1pXWAi8J4s9gxJJ0saFxFzK7Z7N7Ao\nIi6uaMejeXKamZlZ85q5S+kyImJBRPw6Ii4DNpOUd5DmaGB+RLxUsawXqD4JtjnQJ+lcST2SZkja\nOGdOMzMza1JhRUWVnYDjWxS734pZnuMiYgLwR3z6w8zMbMg0NU13i8wDNpS0QsXRijGkoxWVeoGZ\nEXFP9vos4CeSRkTE4lqBp0yZwsiRI5dZNmnSJCZNmlRc683MzErq3HPP5dxzz11m2YIFC+p+f8cV\nFRHxqKQe4ABguqS9gXmV4ykylwPflfTqiPgX8EFg1kAFBcDUqVOZMMGjhM3MzGqp9UW7p6eHiRMn\n1vX+jisqMocDZ2bjMhaQZvBE0nHAQxFxWkQ8K+lw4LeSyLb72BC118ysMJ4p0sqqI4uKiJgDbFNj\n+TFVr68EtmpXu8zMWs0zRXYOF3eNq7uokPSjBuJunaMtZmbDnmeK7Awu7vJp5EjFZxuMHQ1ub2Zm\nS3imyKHk4i6fuouKiGjV5admZmYdysVdI1womJmZWSFcVJiZmVkhXFSYmZlZITryklIzs07mSw3N\nanNRYWbWAF9qaDYwFxVmZg3wpYZmA2u6qJC0IbAecF9EPNN8k8zMysCXGppVyz1QU9Keku4B/gn0\nAG/Plo+SNFPShwpqo5mZmZVArqJC0u7AxUAfcByg/nUR0Qc8BBxURAPNzMysHPIeqTgauD4itgV+\nUmP9X/CNvszMzIaVvEXFFsAFg6x/mDTOwszMzIaJvEXFs8Dqg6wfBzyWM7aZmZmVUN6rP64BDpT0\ng+oVkjYAPgVc1kzDzMzMul23TaSWt6j4GvBX4G/AhaTbnL9P0s7AYaSBm8cV0kIzswZ0207aulc3\nTqSWq6iIiNmStgV+CBxPKiK+lK2+FvhMRDxQRAPNzOrVjTtp617dOJFa7smvIuIuYBdJawGbkMZn\nzI2IR4tqnJlZI7pxJ23DQfdMpJarqJC0eUTcDRART5BOg5iZdYju2UmblUneIxV3SroTOA+4ICLu\nK7BNZmZmVoB2jzHKW1QcAewDfBM4XtKtLC0wHszdGjMzMyvEUIwxyjtQcxowTdL6wEdJBcZ3gO9I\nuplUYFwYEf/K1Soz61q+OsOsPYZijFFTdymNiIeBk4GTJb2GpQXG/wAnASs1E9/MuouvzjAbCu0b\nY9T0rc8rzAfuAmaRpvEebMZNMxuGfHWGWXdrqqiQJGBHYF/gw8Ao4AnS6Y/zm22cmXUrX51h1o3y\nXlK6Hek0x96kG4f9B/gVqZC4MiJeLKyFZmZmVgp5j1RcBzwNXEoqJH4fEc8X1iozMzMrnbxFxUeB\n30bEoiIbY2ZmZuWV95LSGUU3xMzMzMqtrqJC0tGkO5GeEBEvZa+XJyLi+KZaZ2ZmZqVR75GKY0lF\nxXeB57PXyxOkO5iamZnZMFBXURERKwz22szMzKzh4kDSypL2kPSmVjTIzMzMyinPEYfngQuBbQpu\ni5mZmZVYw0VFRARwL2n2TDMzMzMg/zwV3wb+V9KFETG7yAYBSNoEmE4qXJ4EJkfErKptNgL+AdwO\niDQwdK+IuL/o9pgNF43eQRR8F1EzWypvUfEO4DHgTknXAg8AC6u2iYg4Kmf8acCpEXGWpL1IBcbb\namz3n4jwDQTMCpDnDqLgu4ia2VJ5i4rPVjx/9wDbBNBwUSFpXWAi8B5IE21JOlnSuIiYW715o/HN\nrLbG7yAKvouomVXKO6NmKy8pHQ3Mj4iXKpb1AmOA6qJiNUl/IxUXvwa+lY35qGnWrIHWmNmsWasC\nW5HuHlpvUZHek97bSI76tu+0+O3I4T7Ux31oX/xGPjs1yGdw7TdIqwCHArdGxPUNvbm++BOAX0bE\n+IplNwFfiYhrK5atBIyMiD5JawIXAH+MiJMGiHkLbA+MrFo7KXuYmZkNd+dmj0oLgOsBJkZEz2Dv\nbvhIRUQskvRd4Mj+LAWbB2woaYWKoxVjSEcrKtvxAtCXPX9S0umk6uBlRUW/s8+eyvjxHoJhVsus\nWbPYf//9gF/SyOkP2I+zz/4l48cv/z2N5+is+O3I4T64D0XlyBf/+y+LP2tWD/vvP7GO9+cfU3En\nMDbnewcVEY9K6gEOAKZL2huYVz2eIht78UREvChpZeAjwMzBYo8fDxNcU5gNYCHpv1D1mOvlv2f8\n+IV1/t9qNEenxW9HDvehkfe4D0MZ/+Xyjo34GnCYpF1yvn95Ds/izwa+DEwGkHScpEOzbbYFZkqa\nCfwdmA+c0KL2mJmZ2XI0c/XH48AfJN0P3E/tS0r3zBM8IuZQY8bOiDim4vklwCV54puZmVnx8hYV\nbyJdMtoLjAA2KaxFZmZmVkp5LykdW3A7zKwOjc546dkuzayd8h6pGJSkEcCuEXFZK+KbDUd5Zrz0\nbJdm1k6FFhWStgH2Az4KrEM6NWJmBWh8xkvPdmlm7dV0USFpPKmQ+DiwEfAM8Afg0mZjm1kt40mz\nXpqZdZZcRYWkV5MmmtoPeDPpyo9Vga8D/xMRzxfWQjMzMyuFuuepkLSGpIMlXQU8CBwD3A3sAbyV\ndP+N2S4ozMzMhqdGjlT8O/v5O9KpjksjYhGApNcV3TAzMzMrl0Zm1FwFeII00dXc/oLCzMzMDBor\nKjYHzgD2Am6WNCebNnuz1jTNzMzMyqTuoiIi7omIr0fEOGAH4Crg08BdpKs9gnQZqZmZmQ1DuW4o\nFhE3RsQRwIbAh4Ee4DngVEn3SjpJ0o7FNdPMzMw6Xd67lAIQES9GxG8iYh9gfeAQ4AHg86QjGWZm\nZjZMFDajZkQ8BZwJnJnNY7FvUbHNzMys87Xk3h8R8S9gaitim3WqRm/2Bb7hl5l1l5YUFWbDTZ6b\nfYFv+GVm3cVFhVkBGr/ZF/iGX2bWbVxUmBXKN/sys+Grqas/zMzMzPrlKioknS7p7YOsf5uk0/M3\ny8zMzMom75GKycBgNxHbGDgwZ2wzMzMroVad/ng1sLBFsc3MzKwD1T1QU9KewJ4Viw6VtEuNTdcE\ndgH+1mTbzArV6DwSnkPCzKwxjVz9sTnw0ex5AG8HJlZtE8AzwPXAF5punVlB8swj4TkkzMwaU3dR\nEREnAicCSHoJOCQizmlVw8yK1Pg8Ep5DwsysUbnmqYgIX4pqJeV5JMzMWsXFgZmZmRUid1EhaVdJ\nV0h6TNKLkhZXP4psqJmZmXW2vJNf7QVcBqwPnJfFOTd7vhC4HfhmQW00MzOzEsh7pOL/ATcDWwHH\nZMtOj4j9gC2ADYH7m2+emZmZlUXeomJz4LyIWAy8mC1bCSAiHgBOAb7SdOvMzMysNPLepfRZ4HmA\niHhS0nOkoxP9HiZN1W1Wl0YnpgJPTmVm1mnyFhWzSUcr+t0KHCDp7Czmx4HeJttmw0SeianAk1OZ\nmXWavEXFJcCRkr4YEc8BJwC/Bp4kzaq5OnBwMU20btf4xFTgyanMzDpP3smvTgJOqnh9maQdgY8A\ni4HfRsQ1hbTQhhFPTGVmVmZ5j1S8TETcANxQRCxJmwDTgVGkox+TI2LWINufCXwCWDMi/lNEG8zM\nzKwxnTqj5jTg1IjYFPgeqcCoSdKHSYNGo01tMzMzsxryTn4lSYdJullSX63ZNCW9uPxINWOvS7r7\n6S8BImIGMFrSuBrbrk+aM2MKoDz5zMzMrBh5T398j3Rr81tJo+ueKKxFMBqYHxEvVSzrBcYAc6u2\nPQ34UkQ8I7mmMDMzG0p5i4oDgRkRsU+RjWmEpEOAByPiunrfM2XKFEaOHLnMskmTJjFp0qSim2dm\nZlZK1Z+VCxYsqPu9eYuKVYErc753eeYBG0paoeJoxRhePu/FTsB2knZj6amP2yXtGRG31Qo8depU\nJkzw1QVmZmYDqf6s7OnpYeLEiXW9N+9AzauAt+Z876Ai4lGgBzgAQNLewLyImFu13f4RsVFEjIuI\n/tk7txyooDAzM7PWyltUfBp4h6SvSlqnyAZlDgcOkzQb+DIwGUDScZIOHeA9gQdrmpmZDZm6Tn9I\neoqXX7K5InA8cLykRaRJrypFRIwkh4iYA2xTY/kxNTbvXzciTy4zMzMrRr1jKmbgeSDMzMxsEHUV\nFRExucXtsA7X6F1EfQdRM7PhJ9fVH5KOBi6OiDsHWP9GYK+I+GYzjbPOkOcuor6DqJnZ8JP3ktJj\ngfuAmkUFsAVwDOCiogs0fhdR30HUzGw4KuyGYlXWJt2Pw7qK7yJqZmYDq7uokLQ9sGPFoo9kdxOt\ntiawL3BHc00zMzOzMmnkSMVOpFMakK4E+Uj2qOVu4HNNtMvMzMxKppGi4nvAyaQJph4hTVA1o2qb\nAJ6NiEXFNM/MzMzKou6iIiIWAgsBJG0MPBoR9V8OYGZmZl0t10DNiHiw6IaYmZlZueW994eZmZnZ\nMlxUmJmZWSFcVJiZmVkhXFSYmZlZIQotKiSNk1TPPM5mZmbWZXIVFZKOlHRe1bIzgHuBOyX9XdJ6\nRTTQzMzMyiHvkYpPAg/3v5D0PuBA4DTSTJrjWDr7ppmZmQ0DeW8othEwq+L1PsD9EXEEgKQNgAOa\nbJvVqbe3l76+vobeM2rUKN9B1MzMCpW3qFDV6/cCv654/QCwQc7Y1oDe3l423XR8dmvy+q2yymrM\nnj3LhYWZmRUmb1ExB/gwcGp26uPVwOUV618LPNlk26wOfX19WUFxNunW5PWYxaJF+9PX1+eiwszM\nCpO3qDgJOEfSE8DqpFMhf6hYvzNwa5Nts4aMByYMdSPMzGwYy3vvj/MkPQZ8gHRE4pSIeBFA0trA\n46SvzmZmZjZM5D1SQURcAVxRY/njwEeaaZSZmZmVT955KhZL+vgg6/eVtDh/s8zMzKxs8s5TUX31\nR7URQOSMbWZmZiXUzDTdNYsGSWsA7wMamzjBzMzMSq3uokLSMdlpj8WkguLs/teVD+AJ0sRX5w0a\n0MzMzLpKIwM1bwZOIZ36+DRpkOacqm0CeAa4Bbi4iAaamZlZOdRdVETE5WQTXElaHTg1Im5qVcPM\nzMysXPLOU3FQ0Q0xMzOzcss9T4WkEaQBmeOAtXj5FSEREcc30TYzMzMrkVxFhaStgRmke3wMdHlp\nAC4qzMzMhom8l5SeAqwKfAhYOyJWqPEYUVwzzczMrNPlPf3xJuBrEXFpkY0xMzOz8sp7pOKfLH9W\nTTMzMxtG8hYV3wU+lc2eWThJm0j6k6TZkm6SNL7GNmMl/V1Sj6Q7JJ0vaWQr2mNmZmbLV9fpD0lf\nqLH4aeA+SecB84DqG4hFREzN2a5ppHkwzpK0FzAdeFvVNg8B74qI57I2/gA4FpiSM6eZmZk1od4x\nFScNsu6zAywPoOGiQtK6wETgPQARMUPSyZLGRcTcJcEjXqh4zwhgdeCpRvOZmZlZMeotKjZuaSuW\nNRqYHxEvVSzrBcYAcys3lLQSafrwMcDtwB7taqSZmZktq66iIiIebHVD8siOVmwlaUXgx8DhwPeH\ntlVmZmbDU+4ZNVtoHrChpBUqjlaMIR2tqCkiXpR0JnAagxQVU6ZMYeTIZcdyTpo0iUmTJjXdaDMz\ns25Q/Vm5YMGCut+bd0bN+0ljJgYSwCLSpafXANMi4ol6YkfEo5J6SLdPny5pb2Be5XiKrA1jgEcj\nYqEkAR8lnQIZ0NSpU5kwYUI9zTAzMxuWqj8re3p6mDhxYl3vzXtJ6XWkqz/GkgZHzsweT2XLngbu\nBtYDvg3cIamRcRmHA4dJmg18GZgMIOk4SYdm27wJ+KukW4HbgFHAkTn7Y2ZmZk3Ke/rjV8DuwA4R\ncUPlCkk7kO4L8o2I+I2kHYHfAicCH6sneETMAbapsfyYiueXAZflbL+ZmZkVLO+Rim8CP64uKAAi\n4jrgJ8AJ2etrSfNO7JIzl5mZmZVA3qLi9cBgYyQeBzapeD2LNI+EmZmZdam8RcVc4EBJq1avkLQa\ncBDLzinxauDRnLnMzMysBPKOqTgWOA+4R9J04B/Z8k2ATwCvASbBktku9wf+1FRLS6y3t5e+vr66\ntx81ahRjxoxpYYvMzMyKl6uoiIgLJT1LGnz59arVdwKfyQZSQrqb6S4MfrpkyDT6gQ+Nfej39vay\n6abjWbTo2brjr7LKasyePcuFhZmZlUruya8i4rfAbyVtCGyULX4wIuZXbfci0JEzcub5wIfGPvT7\n+vqy+GcDL7vZag2zWLRof/r6+lxUmJlZqTQ9o2ZWRMxf7oYdqPEPfMj/oT8e8MRbZmbWveq99fkn\nsqdnRURUvB5URPwid8vayh/4ZmZmzar3SMWZpKm3zwOez14vTwAlKSrMzMysWQ3d+jwinq98bWZm\nZtYv163PO/VW6GZmZjZ08k5+ZWZmZraMegdqXp0jdkTEu3O8z8zMzEqo3jEVK5AGXjZCDW5vZmZm\nJVbvmIodW9wOMzMzKzmPqTAzM7NC1F1USDpH0jYVryVpjKRXtKZpZmZmViaNHKn4GDC24vXawP3A\ntkU2yMzMzMqp2dMfHoxpZmZmgMdUmJmZWUFcVJiZmVkhGr31+daSFmXPX0Wau2JbSWvW2jgiLm6m\ncWZmZlYejRYVn88elY4dYNsARjTaIDMzMyunRoqKnVrWCjMzMyu9uouKiLiulQ0xMzOzcvNATTMz\nMyuEiwozMzMrhIsKMzMzK4SLCjMzMytEXUWFpDdJGtnqxpiZmVl51XukYibwwf4Xkq6W9O7WNMnM\nzMzKqN6iYiGwWsXrHYH1C2+NmZmZlVa981TcBnxB0mJgQbbsrRVTdtfkabrNzMyGj3qLiqOAi4Cf\nZ68jW3ZaTVTFAAAgAElEQVTUIO/xNN1mZmbDSF1FRUT8XdImwOtIpz2uBU4Armxd08zMzKxMGpmm\n+0VgNjBb0nTgsoi4qRWNygqY6cAo4ElgckTMqtpmC+AnwLrAi8DNwGci4rlWtMnMzMwGl2ueiog4\nqLKgkLSqpFWLaxbTgFMjYlPge6QCo9oiUhGxOfBm4JXAVwpsg5mZmTUg9+RXksZIOkPSw8DTwNOS\nHpZ0uqSNmoi7LjAR+CVARMwARksaV7ldRNwXEXdmzwP4GzA2b14zMzNrTiO3Pl9C0mbAjcCawBVA\n/6mJzYBPALtL2jYiZucIPxqYHxEvVSzrBcYAcwdoz+rAJ/GRCjMzsyGTq6gAvgO8BGwVEXdUrsjG\nOlyVbfPh5pq3fJJWAs4Dfh8Rv2l1PjMzM6stb1GxA/A/1QUFQETcKelk4As5Y88DNpS0QsXRijGk\noxXLkLQicD7wUERMWV7gKVOmMHLk0tnGFyxYMMjWZmZmw08zn5V5i4qVSLNsDuTZbJuGRcSjknqA\nA4DpkvYG5kXEMqc+JI0gFRSPRcTh9cSeOnUqEyZMWPK6p6eHiRMn5mmmmZlZV2rmszLvQM2ZwCdr\n3WRM0hrAIUBPztgAhwOHSZoNfBmYnMU+TtKh2Tb7Ah8CtpY0U1KPpB83kdPMzMyakPdIxTHA74F7\nJJ0BzMmWbwocCKwDfCZvoyJiDrBNjeXHVDw/Bzgnbw4zMzMrVq6iIiKulvQB4PvAf1etvhU4ICKu\nabZxZmZmVh55j1QQEVcCW0naAOifl+LBiPh3IS0zMzOzUsldVPTLiggXEmZmZsNc7hk1zczMzCq5\nqDAzM7NCuKgwMzOzQrioMDMzs0LkKiokbZ/dTXSg9aMkbZ+/WWZmZlY2eY9UXAO8Z5D17862MTMz\ns2Eib1Gh5axfGVicM7aZmZmVUN3zVEgaA4ytWLTZAKc41gQOAx5srmlmZmZWJo1MfnUQ6Z4fkT2+\nlj2qiXSU4rCmW2dmZmal0UhRcQFwJ6louAD4EXBD1TYBPAPcGhEPF9JCMzMzK4W6i4qImAXMApB0\nEHB9RNzfqoaZmZlZueS9S+n0ohtiZmZm5Zb7hmKSxpPGWYwD1uLlV4RERLy7ibaZmZlZieQqKiQd\nAJwBvADMBp6otVkT7TIzM7OSyXuk4lhgJrBrRPQV1xwzMzMrq7yTX70aON0FhZmZmfXLW1TcTios\nzMzMzID8RcUXgEMkbVNkY8zMzKy88o6p+AqwALhB0t1ALy+/10dExJ7NNM7MzMzKI29R8SbS7Jm9\nwCuBzWtsE3kbZWZmZuWTd/KrsQW3w8zMzEou75gKMzMzs2XkLiokjZD0MUnTJF0iacts+UhJH5G0\nfnHNNDMzs06Xq6iQtCbwJ+AcYBKwB7Butvpp0h1MjyqigWZmZlYOeY9UfAd4I/A+0r0/lkzJHRGL\ngYuADzTdOjMzMyuNvEXFh4AfR8QV1L7KYw4wNm+jzMzMrHzyFhUjgfsHWb8STdwB1czMzMonb1Hx\nD2DCIOvfC9ydM7aZmZmVUN6i4v+AgyXty9LxFCFpZUknAO8HphXRQDMzMyuHvKcofkgaqHku8GS2\n7BxgnSzmtIj4efPNMzMzs7LIO6NmAJ+SNB3YG3g96ajHP4ALIuL64ppoZmZmZdDUYMqIuBG4saC2\nmJmZWYnlnfxqY0m7D7J+d0lj8zbKzMzMyifvQM2TgCMHWf8Z0gRZuUjaRNKfJM2WdJOk8TW2WV3S\n7yU9KunxvLnMzMysGHmLincCVwyy/ipgu5yxIV05cmpEbAp8D5heY5sXSIXLu5vIY2ZmZgXJW1Ss\nBTw1yPqnSVeCNEzSusBE4JcAETEDGC1pXOV2EfF8RFwLLMiTx8zMzIqVt6joBd41yPrtgH/mjD0a\nmB8RL1XlG5MznpmZmbVB3qs/zgW+Ielm4OT+AkDSCOCzwL7ACcU0sThTpkxh5MiRS14vWOCDHGZm\nZpWa+azMW1ScCGwL/AD4mqTZ2fJNSbdAv5b8RcU8YENJK1QcrRhDOlrRlKlTpzJhwtLZxXt6epg4\ncWKzYc3MzLpGM5+VuU5/RMRzpPt7HALcDIzKHjcDBwO7ZNvkif0o0AMcACBpb2BeRMwd4C2i4tbr\nZmZmNjQaPlIhaSVgPPB4RJwBnFF4q+Bw4ExJXyUNxJyc5T4OeCgiTste30YqZl4lqRe4JiIObEF7\nzMzMbDnynP54CbgF+C/gR8U2J4mIOcA2NZYfU/X6za3Ib2ZmZo1r+PRHRCwGHgRWLr45ZmZmVlZ5\nLyn9MXCopLWLbIyZmZmVV96rP0YAzwH/kHQR8ACwsGqbiIipTbTNzMzMSiRvUXFSxfNDBtgmABcV\nZmZmw0TeomLjQlthZmZmpZerqIiIB4tuiJmZmZVb3iMVAEh6DbA9sB4wIyL+mU3VPRJYkF0pYmZm\nZsNArqs/lPwvcD/pbqL/C7whW/1K0sDNzxXRQDMzMyuHvJeUfgk4ijRg8z1UTJMdEQuAi4G9mm6d\nmZmZlUbeouJTwC8i4qvArTXW387SIxdmZmY2DOQtKkYDfx5k/TPAGjljm5mZWQnlLSoeIRUWA5lI\nAbcqNzMzs/LIW1RcDBwuaVzFsgCQ9F7SXUUvbK5pZmZmViZ5i4pjgPmk8RS/IBUUX5F0I3A5aUzF\ntwtpoZmZmZVCrqIiu8LjHcD3gNcAi4AdgDWB44DtIuLZohppZmZmnS/35FcRsRD4VvYwMzOzYa6h\nokLSKsCepHt/PAZcFhHzW9EwMzMzK5e6iwpJ65EuI92YpZNdPSvpQxFxZSsaZ2ZmZuXRyJiKbwBj\nSbcz3w34PLAQmFZ8s8zMzKxsGjn98V7SLJpf7F8g6WHgHEmbRsTswltnZmZmpdHIkYoxwI1Vy24k\nnQpZv7AWmZmZWSk1UlSsTLp0tFL/66ZuoW5mZmbl12gxMFbShIrXI7Ofr5f0ZPXGEdGTu2VmZmZW\nKo0WFcdnj2qnVL0WaZbNEXkaZWZmZuXTSFFxUMtaYWZmZqVXd1EREdNb2RAzMzMrt7w3FDMzMzNb\nhosKMzMzK4SLCjMzMyuEiwozMzMrhIsKMzMzK4SLCjMzMyuEiwozMzMrhIsKMzMzK4SLikGdW/L4\n7cjhPgx9/HbkcB86I4f70Bk53IeBdGRRIWkTSX+SNFvSTZLGD7DdbpJmZdtdJOmVxbaknP+o7c3h\nPgx9/HbkcB86I4f70Bk53IeBdGRRAUwDTo2ITYHvAS+bIlzS6sD/AXtk280Hjm5rK83MzGyJjisq\nJK0LTAR+CRARM4DRksZVbbor0BMR92avTwEmta2hZmZmtoyOKyqA0cD8iHipYlkvMKZquzHAgxWv\nHwA2kNSJfTIzM+t6jdz6vMxWAZg1a9YyC5e+/h0wi5f7J9kBkwr3V713cIPnaHX8duRwH/LFb0cO\n96Ho+O3I4T7Ux31oX/yK16ssL6cioq7GtUt2+uNeYO3+oxWS5gPvioi5FdvtDRwSEbtmrzcHfh8R\n1Uc0kPRxau8RzczMrD77RcQ5g23QcUcqIuJRST3AAcD0rHiYV1lQZH4PnCzpDRExBzgCOG+AsH8A\n9iOdIlnUmpabmZl1pVWAsaTP0kF13JEKAElvAM4E1gEWAJMj4m5JxwEPRcRp2Xa7Ad8HRgB3AgdG\nxFND02ozM7PhrSOLCjMzMysfXylhZmZmhXBRYWZmZoVwUWFmZmaF6LirP4aSpLWAjYAXgfsiovAr\nRVqdw30ws+GoG/ZL3dAHImLYP4DXApcBi7PHY8CzpPuOrFSGHO5Dw7nWAt4CbAGs0qK/q5bmcB86\nI0fZ45e9D92wX+qGPvQ/fPojOYN0y7ZRwBTgR8DGwHqkS1bLkMN9qIOk10q6DOgDbgGuAx6X9D1J\nK5Uhh/vQGTnKHr8dOdrRB7pgv9SGHO3oQ9KKqrdsD+D2qtc3Zz9HAPeWIYf7UHeOK0gToa0FHAkc\nA6xPmhflB2XI4T50Ro6yx++iPnTDfqn0fVgSu8hgZX0AtwLrZc83Af5Sse6uMuRwH+rO4R1EB+Rw\nH4Y+fhf1oRv2S6XvQ//DAzWT/wFuk3QbMAE4HEDSBix7J9ROzuE+1OclSetFxCOSNiGdXyQiFkt6\nviQ53IfOyFH2+O3I0Y4+dMN+qRv6kBRZoZT5AWwK7AVsUtYc7kNd8Q8A5pPuHfMI8JFs+QbA78qQ\nw33ojBxlj98tfcjilXq/1C19iAhP023Dj6RNSaPQb4uI+8qYw33ojBxlj9+OHO3og3UOFxWApI0j\n4v7suYD/ArYDZgInRMQLnZ7DfTCz4agb9kvd0Id+vqQ0mVHx/OvAe0mX32wOnFSSHO5DHSRtXPFc\nkr4o6deSji3wMr2W5nAfOiNH2eO3I0c7+kAX7JfakKMdfUhaeW6lLA9gZsXzvwNrZs9XBu4oQw73\noe4cPRXPvwH8EfgYcAHwwzLkcB86I0fZ43dRH7phv1T6PvQ/fPVHUnkOKCLiyezJc5JeLEkO96E+\nqni+J7BLRDwp6RLSf7Yy5HAfOiNH2eO3I0c7+tAN+6Vu6APge3/0e5Okx0n/AVaTNCoi+iStSHG/\no1bncB/q4x1EZ+RwH4Y+fjtytKMP3bBf6oY+QNHBSux1Va//k/1cCzi6JDmq4y8oOH6tHK3uQyv+\nHbyD6Iwc7sPQx29Hjnb0oRv3rUXnaMe+FfDVHzbMSNqoatH8iHhe0rrAthFxSQty/CsiXigqR6vj\nD5Cj0N9Tl/bBv6M2x7fO46ICkLRyRDxX8XpvssttIuLMgnJ8CLghIh6TNAr4KfB24Dbg8Ih4qMn4\nFwFnA5dGxOKmG1w7xzjg28A/SdXtNGBH4E7giIh4oMn4KwKfB/YHxpJuz3sX8N2I+F0zsa3zSVon\nIh4b6nZYuUgaARwKfBwYnS3uBc4DprVqf5jlPiEivlZQrE+Q9nuXRURPxfL/FxEnNhl7JeDTpNNR\npwAfId2T5Xbg+IgoanZTX1Ka+Uv/E0lHkEYpzwc+KekbBeU4AXg8e34ScC/wfuB64LQC4u8AHA88\nJOn7kjYrIGa1nwF3kKbavQ74F/A+4Frg1ALi/xR4A3AccDlwMuk/wAmSPlVAfCSNkHSEpBskPZA9\nrpf06Wzn1FKSTigozickHS1pQtXy/1dQ/JUkHSXpSEkrStpH6VLA4yW9oogcNfytyGCSVq56vbek\nH0qaXFD8D0laJ3s+StKFknolXSrpNQXEvyjL0bK/S0njJJ0n6SRJq0k6S9I8SZdLGltA/BWVLiO9\nVdKTkvokXSfpA823folTSPvSbwO7Zo8TSfumnxaYp5YDiggi6bvAp4B1gd9KOqpi9UcLSPEj0mfE\nrsBZpCtwzgFeT7r9eWF8pAKQNDMitsqe30SaSvYhSa8C/hwRWxaQ4+6I2Lw6X/b61oh4S5PxZ0bE\nVpLeDhxM+qO5E/g5cH5EPNNM/CzHbRHxZkkCHoqIV1esK6IPlb+jlYBrImLb7MjOdRHxxqY6kOJO\nI00RfCrwQLZ4LGku/Icj4tBmcywnf29EjGkyxneBbUg3Cdob+E5E/DBb1xMREwZ7f505fkq6m+Sq\nwJOkS8/OBz4M/DsiPt9k/CdYdhAfwEjSud6IiLWbiZ/lWPK7yL4sHE66Nn834A8RcXyT8e8CtoiI\nkHQmqcg+G/ggsGNEfLDJ+I8C/yZ90JwF/Dwi7mkmZo0cVwFXA2sAO2fPpwO7AztFxPubjP8z0r/z\n5cA+wGxgFvDfwCkR8bNm4mc57o2I19dYLmBOrXUNxu8ZaBUwPiJWaSZ+luMOYGJ2amgD4DfAJRFx\nYvXnRd74EbGlpFVI06VvEBHPZl8QbiniM66fB2omlTu3lfpPRUTEUypuhPLDkt4ZEX8B/ilpg4j4\nd1a4FPFNJAAi4ibgJklTSP+JDwamknbYzVpR0grAasAakl6V/Y5WAor49rpY0ojscOUqWR6ygV0v\nFRAfYOcaO5lZkn4PzCkiwXJ2QusVkOIDLN0BnQD8RtJq2SFSLee99dp2gJ3QJcAtBcSfSTpa9x3g\nJVK7bwC2LSB2v8rfxWTgA9mXhZ8AfyYd2Wsqfiz9VvbmiJicPb9b0n5Nxgb4Z9UXhZskFfpFARgV\nESdUfFH4Srb8bkmTCoj/roovCpeRvigcmxUz15GOfjYrJK0bEY9WLV+XYv4/jAMmAc9WLRep0C6C\n+k9BZJ8LuwCXZ0epivjm/0IWe5GkuRHxbPb6+QI/4wAXFf02yz4IBLyu4sNSQFGzvh0J/ErSX4BH\nSTuIa4C3kg7bNWuZ/zzZH82ZwJmS3lBAfEg3BfoT6Vvrz4AZkq4mfcO5sYD4fwD+IOlKYA/gVwCS\n1qSYwgtavwOC1u+EWr0DghbvhCJi56zwPQM4NCLmSHohIoq8Y2Krvyz4i8LyteOLwneBWyX9mqV3\n3NyItA85poD4M4EFEfHn6hUq7k6rz0ga2z8uLSL+I+l9pH1i00doSQdu+v8d9qxY6EtKW2TXqtf9\nO6P1KeicXETcIWkL0mCizUkfmL3ANyJiXgEpfjhI7kK+gQNfJP1HDeBSUkF0AOnQ5k8KiP8l4BDg\nzcBpFYNkF5IGhBah1TsgaP1OqNU7IGjDTigipkr6I3CGpBkUV9T1a/WXBX9RWL6Wf1GIiJ9nbd4L\n6D+1OBfYLrL7XTRpMksvwaxW1L/Df1NVJEbE05LeCxxV+y0NOYL0N7+4qnDfiDTeojAeU2HDjtIA\ntL1ZugPqBWYUtAPqv4zuPxHxRI11q0bEwibj7wQ8HhG3VS1fHTgqIpr+QJP0DuDWiFhUtfx1pFNI\nRRy27o+5Iqmg2yEiti8w7g5Vi27JdtQbAHtFRNOFsKRVWfpFYUXS39IFRXxRkDQ5Crr6bJAcYtkv\nClsDnyB9KP+k2asCsvj9XxRu6e+P0iDakRHxSDPxrfO4qMhI2pV0yLryg+a8KPBSxlbnGMI+nBsR\nl7cwfqF9MLPuIml70oDrKyJifsXyAyNieoHxr4yIfxUdvx052tEHwDcUy4qqb5EuZzuCpZckHZEt\n+1YZcrgPhfTv8LLncB86I0fZ45epD8AXSIN+LyWdhvpIxbqeTo/fLX1YEq+VfzRleWS/7FfUWL4y\ncF8ZcrgPhfSvt+w53IfOyFH2+GXqA2kCpzWy55uTruLaP3s9s9Pjd0sf+h8eqJmI2hOBrUBxg8da\nncN9qCeBdPEgudcpQw73oTNylD1+O3K0ow+QBitnP++WtDNwRZFXQ7U6fjtytKMP4Ks/+p0J/E3S\nWSx7RcABpEveypCj1fHbkaPV8SHNsvd5oHoAmkhTs5chh/vQGTnKHr8dOdrRh8WS1ots0GdE/FPS\nu4ErWTptdyfHb0eOdvQhKfKwR5kfwPakaaF/kz1OJo1GL00O96Gu+H8Cth5g3bwy5HAfOiNH2eN3\nUR/2Bd5ZY/mGpEvTOzp+t/Sh/+GrP2xYkbQ16W6P/6qx7g1RwJwerc7hPnRGjrLHb0eOdvTBOouL\nigFI+ktEvLPMOdwHMxuOumG/VNY++C6lA2v6JjEdkMN9qEM2I2Kpc7gPnZGj7PHbkaMdfaAL9ktt\nyNGS+C4qBlZ934Yy5nAf6uMdRGfkcB+GPn47crSjD92wXyplH1xUDCAi3lX2HO5D3byD6Iwc7sPQ\nx29Hjpb3oRv2S2Xtg8dULIek3SLisoJijeflU1CfHxF3lSF+O3K0ow9m1l26Yb/UDX0AFxXLJak3\nIsYsf8vlxvkM6S6c5wMPZIvHki71OSkiTu7k+O3I0Y4+ZHm8g+iAHO7D0MdvR442xC/9fqkb+rAk\nl4sKkPS/A60CDo6IkQOsbyTHHODtUXXnSklrAzdFxOs7OX47crSpD95BdEAO92Ho47cjR5v60A37\npdL3YUlMFxUgaRHwPWBxjdVTImLNAnLcFxGb1Fi+AnBvRLyuk+O3I0eb+uAdRAfkcB+GPn47crSp\nD92wXyp9H/p5mu7kTuDCiLijeoWkTxaU43JJVwA/Y9kpqD8FFHFb71bHb0eOdvRhheodXOZJihu4\n3Ooc7kNn5Ch7/HbkaEcfumG/1A19AHykAgBJHwDmRMR9NdbtHBFXF5BDpHtY7MOy5xYvBM6KiJc6\nOX47crSpDz8GNqP2f657IuJznZ7DfeiMHGWP344cbepDN+yXSt+HJblcVNhw4h1EZ+RwH4Y+fjty\ntPPDzDqDi4qMpLWAD7PsH/6vIuLxsuRwH8xsOOqG/VI39AFcVAAgaS/gFOAalh6iGwvsAHwmImYU\nmONalh0FXUiOVsdvR4529CHL4x1EB+RwH4Y+fjtytCF+6fdL3dCHJblcVICke4D3R8QDVcs3Bi6P\niM06PYf7UHcO7yA6IIf7MPTx25GjTX3ohv1S6fuwRBR4H/WyPkiX1Ay07r4y5HAf6s5xDzC2xvKN\nSQPHOj6H+9AZOcoev4v60A37pdL3of/he38kf5N0uqS3SVo/e7xN0unA30uSw32oz4ioqtYBIuJ+\nirvEutU53IfOyFH2+O3I0Y4+dMN+qRv6APj0BwCSVgW+SJrlrf+834PARcD3I6LpG+C0Oof7UHeO\nc4BFwKkse4nb4cBqEfGxTs/hPnRGjrLHb0eONvWhG/ZLpe/DklwuKmw48Q6iM3IMEL//UsMy96E0\n8duRo50fZtYZXFRkJI0gDR6q3MFdFxG1pu4uIt+WwFuB2yOi6cNPkraIiDubb9ly82wGPB4Rj2TP\n3wXcERE3FxR/ZWBX0mCuF4E7I+LaImKbWXdq9/67Iu86EfFYgfFWiKq5OyStFbVnJW0mz4rAlsDc\niFhQZGyPqQAkbUcamfxt0gfarsCJwAOSti8ox1WS1sue7wP8Hng/cJGkwwpIcbuk2yQdqTSvfuEk\nfQm4Dvi7pP2BPwLvAy6UdFQB8XcC5gDHAd8B9gROkXSzpNc0G78izwhJO0uanD12znZKLSdpnQJj\nvez/r9Lle4WTtKKkrSQ1fXO9QXIc36rYFTla8vvJYm8p6WBJWxcUb4si4tSRZ7OKfdNmkg6R9LYC\n468s6UOSPi/ps5J2LCp2Fr/l++9BzCwiiKStJd0PLJR0iaR1K1ZfVUD8nSU9JqlP0g7An4FzgH9k\nr4tT5KjPsj6A24Gtayx/K+lbeCE5Kp7/Bdgoe7525bom4t8B7EWax/1p4Dxgl4J/T3cBawGjgWeA\njbPlo0hHFJqN3wO8vuJ3/4vs+adI17UX0YftgHnAX0l3TjwfuClbtn0b/tZ6C4ixNXA/8BxwCbBu\n5e+woHbuDDwG9JG+Ad4MzOp/XUD8I2s8+vqfF9SHoyqeb5z9/S7MfndbFhD/KmC97Pk+wEPABaQP\nuMMKiP8ScFv2O1m7RX+PXwIeJn2z3z/7eQHpFMVRBcTfKYt1G2lsxRXA3dnf02sK6kNL99/AHoM8\nHimoDzcAHwTWAY7P/q+9Jls3s4D4NwFvAXbM/p/tnC1/G3BDoX9TrfhDLduDdN+Phtc1mGM2aSQ0\nwF+r1hXxh99T8Xw08HXgH9kO7uiC+lCZ48GqdUX84d9a9fqWyt9fQX1oRwHZ0p1Qq3dAWZyW7oRI\np7Z+DZxR8Xgq+3l6QX2o/Hs9lzQvAqTi+4oi/pYqnvuLwgD/BrT+i0JL99+ku1dfTZocsfqxsKA+\nzKx6vX/2mTGaAr4oVMan6hLSovYZ/Q/fpTT5h6SjgVMj4hGA7HDgEaRvNUU4Fzhf0n+TTnl8HTib\ndKhubkE5AIiIecC3gG9JejdwcEGhn5P0QdJOKCTtGxHnZ6ctijh3+bSknSLiGkl7A48UELPaKlFj\nDEtE/C0bz1GES0iniVRj3asKiP/KiPht9vwbkmYDV0vaBShqkNQrIuJWAElPRnZTvYi4WdIrC4j/\nXtIh6p9FxGVZnh0j4qACYteyeURMAoiIGZK+UUDMlSWNiHTeXhHxYBb/cUm1/u0b9UKkyaFmSBoN\nHAhMy07VnR4R3ywgx3ORztc/Iakv0qWeRESfpBcKiL9CRNybxfybpDdmz38m6YsFxIfW77/vBQ6O\nGpfGSppXQHyA1SrHU0TE2dnv/yqgiP1S5anSC6vWFXvqt8gKpawPYF3gdNI3pYXZ46ls2XoF5jmK\ndJj9edKhzQWkS62aPrRJ1dGPFv2e3kY6h3gL8GZSobSIim+yTcZ/a/b7WUgqtLbMlm9AcUdbLgeO\nrvx3BdYDjgH+UFCOmhP+ZOvmFRB/NmlnXblsX9J4lAebjZ/Fu63i+YlV65r+Fp7FWQP4BenoxBqk\nQWNF/r3OBXYnjc25p2rdbQXEP5Z0FcMmpCscvkYaYHwE/P/27j5GrqqM4/j3FyqJgVqgxLaCYLoW\naH2JBhutUaRAldQKJjQYQ3mRmKKAYqomvGhoRUUMMWpAbKG1RUGUBosCAUNfEBQU37ABTKxgWmpB\nUUtpoVbo4x/nzHY63NlOd89Md2Z/n2SyO2fuPM89k3bnueeeew+3F4hfeYQKnAjcVOgzepA06jWH\nNKr5kdw+HfhtgfgPANPz77NJd2/s/3dcqA9t/fsNXELF6GZ+7bJCfVgCzKpoPx3YUSj+ayra+4AH\nSvSh9vDVH3Uk9ZG+LCENxRcdQajLMYV045dHSueQNJG6GdBt7MPr89MtpM+q3NK50rGkL5mgcB/y\nBKivkf6z1kbqXiJV7xdHPtIZYo5LSMPrrxgRkXRZRHxliPGXALdFPsKvaz8d+EFE7D+U+HU5PhMR\nWxra+4BlEfGeoeaoi3ka6XTd+IiYUDDuGnYfuZkTERvzUeydETG1QI6LSAXFONK/p+dJxfalMcS1\nLSQ9FBHvGuo+7iHHVGAR6SDnXOBi0jodW0kFxpAmCeb4t5FOp2wCTo2ItZLGA3OjzGhLfb5DII0W\nlYzbq/JVIKMiYnuxmC4qQNIUYClwOGmikkhfmk8BH4uIR4d7DkmTgWU55vrcfATpyL9UH2o5Ds9x\ni+ZoiF/7jGrxz4mIx4YSvyKf/wDtpXb8EcpxxwHHRsRdJeM2ybUf6fTOiwVjjiZ9LkUv/dsX8hVK\n/180XzIAAAbySURBVCl8oFD00suG2H3ADaSbaq0gFXTb82sPRsS0IcafCCxuV/xO5OhEH/qVHPbo\n1gdpUtppFe2zgd90Qw73oUj/ikzK3Zc53IfhkaPb43dTH4B7gAuAY0mn034JjM6vlZhA3tb4vdKH\n2sMTNZODomK1vIhYLmlIQ9UdzOE+tEDSWwd4ucQkyrbncB+GR45uj9+JHJ3oA2nexLX597MkXQqs\nlDSDMhOX2x2/Ezk60Qeg3IIu3e5ZSWeSJj/thP4bC51Jula/G3K4D635I2lCWtXs/FI3pmp3Dvdh\neOTo9vidyNGJPry6/klEfFXSDtKVEyUKl3bH70SOTvShP/iIf5Bmb68ENpOu+X88/74KOKobcrgP\nLed4Enhdk9eGfGVGJ3K4D8MjR7fH76E+/AQ4uaJ9HrBzuMfvlT7UHh6pACJiHXBivjKgdlXDhoj4\nZ7fkcB9a9lNgIvD3itfurGgbjjnch+GRo9vjdyJHJ/pQudJpRHxD0o+6IH4ncnSiD4Cv/jAzM7NC\nvKCYmZmZFeGiwszMzIpwUWFmZmZFuKgwMzOzIlxUmJmZWREuKsysoyTNl1RsXYl2krQzL6tdez4/\ntx2yL/fLbLhyUWHWAySdnb/sao8XJW2UdLekT0k6sIO5mz1qq80GaVXMjmuyr89IWiXp5BZCBIVv\na2zWS3zzK7PeEcAXSbdFfhUwHjge+CYwT9IpEbG2DXnvA+Y0tC0mLRC3qK5ta/55BXBlG/ajVfWf\nk0jLlp8D3CVpVnRgpVSzXuWiwqy33B0Rv697fpWk40l3L7xd0uSI+G/JhBHxN9IXdD9JC4EnIuLm\niu13AjtK7sMg7PY5SVoCPAN8FHBRYTZIPv1h1uMiYg1pdOBI6kYUJB0tabmkf+XTJQ9L+lD9e+tO\nF7xX0kJJz0p6TtIySQcNZn+q5lTkHN+WNFvSo5JekPQrSW/Or58n6S95P1dLOqIi7jvz6Z7NkrZJ\nWiPp3a3sU0RsBl4EXhpEf46UtE7Sn/It5s1GLBcVZiPD90lD/e8HkPQm4CHgaNKpiHmk0xMrJJ1a\n8f5r8raXA8uAM0iLFA1Gs3kJxwFXA0tznsnAHZLOBy4ErgW+DkwDltS/UdIJpNMwBwLzgUuAMcAq\nSe+oyDVG0lhJh0qaIum7wAGkz6llkvqAX5AWvntf4XVqzLqOT3+YjQARsVHSc0BfbvoW6ZTF1Iio\nHZ1fJ+kB4Crg9oYQ24ETI+JlAEnrSadWZkXEHYV28yjg6IjYkHNsBhYClwGTIuKF3D4KuFjSERGx\nvrbvwMqI+GAtWD4F8xjwZaB+EqZIq+E29u/ciFjV6s5KOga4F9hAWgHyuZZ7atajPFJhNnJsBUZL\nOhiYDtzKriP2sZLGAj8HJkma0PDeRbWCIrsOeBmYWXD/7q0VFNmv88/ltYKioX0igKS3AZOAHzb0\nZTSpeDiuIU8AnwROyo8zgNXAYkkfbnFf3wKsAZ4AZrigMEs8UmE2chxImoz4RtLR+hWko/hGAbwW\n2FT3fN1uG0Rsk7QJeEPB/dvQ8Lz2Rf1URbuAg/PzSfnnjU3i7pQ0puGL/+GGiZq3AH8ArpF0R93o\nTRUBPwOeJo1QvDDAtmYjiosKsxFA0mGkOQbr2DVCeTVwT5O3rGvS3k4v72W78s9afz4LPNJk261N\n2gGIiJC0Gvg0qUh5fKDNgeXA2aSJr4sG2NZsRHFRYTYynEX6MrybNGQP8L8W5xCI9EV7X3+DdAAw\ngXSp6r721/zz+b2ZE1Gh9vewlRuFfZ5U7HxH0paIuGUIec16hudUmPW4fGXEF0jFxM35CoU1wHmS\nxldsf2hFmLl5gmTN+cB+DI97OvyOVFh8Lhc7u2nSn8ZtRgEfIN0/Y6BRipoA5pJGLG6UNGuv9tis\nR3mkwqx3CJgpaTLp//Y44ARgBvAkcEpE1G46dQFwP7BW0vWkgmMc6XLNw4C3N8TeH1gp6cfAMaSJ\njvcXvPJj0PKpi4+TCpxHJX0P2Ejqx3TSHIz6y2TrPydI80fOIF0Zc2VEDHiqpCHvHGAFcKukmRGx\nukinzLqUiwqz3hHAgvz7DuDfwFrSPIGlEbGtf8OIx/P9Gy4nzQ0YC/yDNFnxSxVxLyR98S4g3QL8\nJuCiPezLQGtkNL7WbPuB2nc9ibhP0jTS7bcvIJ3CeJp0pcjCivcuqHu+Hfgz8ImIuL7F/LW8L0ma\nTSpoVkg6KSIebra9Wa9ThNfGMbNqks4m3WhqasPtv83MXsFzKszMzKwIFxVmtifa8yZmZi4qzGzP\nfI7UzFriORVmZmZWhEcqzMzMrAgXFWZmZlaEiwozMzMrwkWFmZmZFeGiwszMzIpwUWFmZmZFuKgw\nMzOzIlxUmJmZWREuKszMzKyI/wNmSGxk/21Q8QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a00187d630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAG7CAYAAAB0Gmd9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxce3XnKwIsAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a001861eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAGRCAYAAADSLUEtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0024731d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAGXCAYAAAAEdKIwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAupAAAWAQv1z5UCGSpcAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a002a2a2e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAGXCAYAAAAEdKIwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nHFSYmZlZKxxUmJmZWSv+Baeiww182qoCAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a002ac1828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Proportion of late flights per category based on all other flights\n",
    "\n",
    "avgLate = np.sum(data['ArrDel15'])/len(data['ArrDel15'])\n",
    "attributes = ['Month','DayOfWeek', 'DayofMonth', 'DepTimeBlk','ArrTimeBlk','UniqueCarrier',\n",
    "             'ArrivalDelayGroups','DepartureDelayGroups']\n",
    "for i,pred in enumerate(attributes):\n",
    "    plt.figure(i, figsize=(15, 5))\n",
    "    group = data.groupby([pred], as_index=False).aggregate(np.mean)[[pred, 'ArrDel15']]\n",
    "    group.sort_values(by=pred, inplace=True)    \n",
    "    group.plot.bar(x=pred, y='ArrDel15')\n",
    "    plt.axhline(y=avgLate, label='Average')\n",
    "    plt.ylabel('Percent of Flights that Arrive Late')\n",
    "    plt.title(pred)\n",
    "    plt.legend().remove()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Exploratory analysis reveals:\n",
    "\n",
    "(1) Delay changes slightly with Month e.g. lower delays in Autumn months       \n",
    "(2) Not much variation within week and day of month (slightly higher delays in the last few days of month)     \n",
    "(3) Higher delays for flights with departure and arrival time blocks from 3pm until midnight    \n",
    "- There seems to be a lag in the pattern for arrival time block        \n",
    "- Pattern starts with departure time block around 2pm and lasts until midnight         \n",
    "- Pattern starts with arrival time block around 4pm and lasts until 1am (with a lag of 1-2hrs)\n",
    "\n",
    "(4)Similar pattern is observed for arrival and departure delay groups          "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Label encoding:\n",
    "Conversion of categorical values to numericals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "le = LabelEncoder()\n",
    "\n",
    "data[\"Unique_Carrier\"] = le.fit_transform(data[\"UniqueCarrier\"])\n",
    "UniqueCarrier = list(le.classes_)\n",
    "data[\"Carrier_Name\"] = le.fit_transform(data[\"Carrier\"])\n",
    "Carrier = list(le.classes_)\n",
    "data[\"Tail_Number\"] = le.fit_transform(data[\"TailNum\"])\n",
    "TailNum = list(le.classes_)\n",
    "\n",
    "data[\"Origin_Point\"] = le.fit_transform(data[\"Origin\"])\n",
    "Origin = list(le.classes_)\n",
    "data[\"Origin_CityName\"] = le.fit_transform(data[\"OriginCityName\"])\n",
    "OriginCityName = list(le.classes_)\n",
    "data[\"Origin_State\"] = le.fit_transform(data[\"OriginState\"])\n",
    "OriginState = list(le.classes_)\n",
    "data[\"OriginState_Name\"] = le.fit_transform(data[\"OriginStateName\"])\n",
    "OriginStateName = list(le.classes_)\n",
    "\n",
    "data[\"Destination\"] = le.fit_transform(data[\"Dest\"])\n",
    "Dest = list(le.classes_)\n",
    "data[\"Dest_CityName\"] = le.fit_transform(data[\"DestCityName\"])\n",
    "DestCityName = list(le.classes_)\n",
    "data[\"Dest_State\"] = le.fit_transform(data[\"DestState\"])\n",
    "DestState = list(le.classes_)\n",
    "data[\"Dest_StateName\"] = le.fit_transform(data[\"DestStateName\"])\n",
    "DestStateName = list(le.classes_)\n",
    "\n",
    "data[\"DepTime_Blk\"] = le.fit_transform(data[\"DepTimeBlk\"])\n",
    "DepTimeBlk = list(le.classes_)\n",
    "data[\"ArrTime_Blk\"] = le.fit_transform(data[\"ArrTimeBlk\"])\n",
    "ArrTimeBlk = list(le.classes_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.drop(['UniqueCarrier','Carrier','TailNum',\n",
    "           'Origin','OriginCityName','OriginState','OriginStateName',\n",
    "          'Dest','DestCityName','DestState','DestStateName',\n",
    "          'DepTimeBlk','ArrTimeBlk'], axis=1, inplace=True)#Removing original encoded columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.drop(['DepartureDelayGroups','ArrivalDelayGroups'], axis=1, inplace=True)#Removing delay details as per the task"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "48\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Year</th>\n",
       "      <th>Quarter</th>\n",
       "      <th>Month</th>\n",
       "      <th>DayofMonth</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>FlightDate</th>\n",
       "      <th>AirlineID</th>\n",
       "      <th>FlightNum</th>\n",
       "      <th>OriginAirportID</th>\n",
       "      <th>OriginAirportSeqID</th>\n",
       "      <th>...</th>\n",
       "      <th>Origin_Point</th>\n",
       "      <th>Origin_CityName</th>\n",
       "      <th>Origin_State</th>\n",
       "      <th>OriginState_Name</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Dest_CityName</th>\n",
       "      <th>Dest_State</th>\n",
       "      <th>Dest_StateName</th>\n",
       "      <th>DepTime_Blk</th>\n",
       "      <th>ArrTime_Blk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.00000</td>\n",
       "      <td>7.338260e+05</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>7.338260e+05</td>\n",
       "      <td>...</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2016.497052</td>\n",
       "      <td>2.476836</td>\n",
       "      <td>5.430509</td>\n",
       "      <td>15.794864</td>\n",
       "      <td>4.02110</td>\n",
       "      <td>2.016553e+07</td>\n",
       "      <td>19904.732889</td>\n",
       "      <td>2072.336574</td>\n",
       "      <td>12691.547570</td>\n",
       "      <td>1.269158e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>156.978973</td>\n",
       "      <td>150.878129</td>\n",
       "      <td>20.642058</td>\n",
       "      <td>20.475231</td>\n",
       "      <td>158.967048</td>\n",
       "      <td>153.688580</td>\n",
       "      <td>20.551899</td>\n",
       "      <td>20.383095</td>\n",
       "      <td>8.590226</td>\n",
       "      <td>10.215711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.499992</td>\n",
       "      <td>1.089987</td>\n",
       "      <td>3.269961</td>\n",
       "      <td>8.905223</td>\n",
       "      <td>2.03702</td>\n",
       "      <td>4.710840e+03</td>\n",
       "      <td>389.869122</td>\n",
       "      <td>1704.295317</td>\n",
       "      <td>1526.501453</td>\n",
       "      <td>1.526499e+05</td>\n",
       "      <td>...</td>\n",
       "      <td>86.031285</td>\n",
       "      <td>89.287731</td>\n",
       "      <td>15.351702</td>\n",
       "      <td>15.325205</td>\n",
       "      <td>86.132155</td>\n",
       "      <td>89.181256</td>\n",
       "      <td>15.385258</td>\n",
       "      <td>15.357587</td>\n",
       "      <td>4.769283</td>\n",
       "      <td>4.945489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>2.016070e+07</td>\n",
       "      <td>19393.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>10135.000000</td>\n",
       "      <td>1.013503e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>2.00000</td>\n",
       "      <td>2.016073e+07</td>\n",
       "      <td>19690.000000</td>\n",
       "      <td>706.000000</td>\n",
       "      <td>11292.000000</td>\n",
       "      <td>1.129202e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>70.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>71.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>6.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>4.00000</td>\n",
       "      <td>2.016103e+07</td>\n",
       "      <td>19805.000000</td>\n",
       "      <td>1613.000000</td>\n",
       "      <td>12892.000000</td>\n",
       "      <td>1.289204e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>170.000000</td>\n",
       "      <td>165.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>170.000000</td>\n",
       "      <td>175.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>11.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>6.00000</td>\n",
       "      <td>2.017013e+07</td>\n",
       "      <td>20304.000000</td>\n",
       "      <td>2878.000000</td>\n",
       "      <td>14027.000000</td>\n",
       "      <td>1.402702e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>229.000000</td>\n",
       "      <td>231.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>230.000000</td>\n",
       "      <td>231.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>14.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>7.00000</td>\n",
       "      <td>2.017043e+07</td>\n",
       "      <td>21171.000000</td>\n",
       "      <td>7439.000000</td>\n",
       "      <td>16218.000000</td>\n",
       "      <td>1.621801e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>314.000000</td>\n",
       "      <td>310.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>314.000000</td>\n",
       "      <td>310.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 48 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Year        Quarter          Month     DayofMonth  \\\n",
       "count  733826.000000  733826.000000  733826.000000  733826.000000   \n",
       "mean     2016.497052       2.476836       5.430509      15.794864   \n",
       "std         0.499992       1.089987       3.269961       8.905223   \n",
       "min      2016.000000       1.000000       1.000000       1.000000   \n",
       "25%      2016.000000       2.000000       4.000000       8.000000   \n",
       "50%      2016.000000       3.000000       7.000000      16.000000   \n",
       "75%      2017.000000       3.000000       7.000000      24.000000   \n",
       "max      2017.000000       4.000000      10.000000      31.000000   \n",
       "\n",
       "          DayOfWeek    FlightDate      AirlineID      FlightNum  \\\n",
       "count  733826.00000  7.338260e+05  733826.000000  733826.000000   \n",
       "mean        4.02110  2.016553e+07   19904.732889    2072.336574   \n",
       "std         2.03702  4.710840e+03     389.869122    1704.295317   \n",
       "min         1.00000  2.016070e+07   19393.000000       1.000000   \n",
       "25%         2.00000  2.016073e+07   19690.000000     706.000000   \n",
       "50%         4.00000  2.016103e+07   19805.000000    1613.000000   \n",
       "75%         6.00000  2.017013e+07   20304.000000    2878.000000   \n",
       "max         7.00000  2.017043e+07   21171.000000    7439.000000   \n",
       "\n",
       "       OriginAirportID  OriginAirportSeqID      ...         Origin_Point  \\\n",
       "count    733826.000000        7.338260e+05      ...        733826.000000   \n",
       "mean      12691.547570        1.269158e+06      ...           156.978973   \n",
       "std        1526.501453        1.526499e+05      ...            86.031285   \n",
       "min       10135.000000        1.013503e+06      ...             0.000000   \n",
       "25%       11292.000000        1.129202e+06      ...            80.000000   \n",
       "50%       12892.000000        1.289204e+06      ...           170.000000   \n",
       "75%       14027.000000        1.402702e+06      ...           229.000000   \n",
       "max       16218.000000        1.621801e+06      ...           314.000000   \n",
       "\n",
       "       Origin_CityName   Origin_State  OriginState_Name    Destination  \\\n",
       "count    733826.000000  733826.000000     733826.000000  733826.000000   \n",
       "mean        150.878129      20.642058         20.475231     158.967048   \n",
       "std          89.287731      15.351702         15.325205      86.132155   \n",
       "min           0.000000       0.000000          0.000000       0.000000   \n",
       "25%          70.000000       7.000000          7.000000      80.000000   \n",
       "50%         165.000000      17.000000         18.000000     170.000000   \n",
       "75%         231.000000      33.000000         33.000000     230.000000   \n",
       "max         310.000000      51.000000         51.000000     314.000000   \n",
       "\n",
       "       Dest_CityName     Dest_State  Dest_StateName    DepTime_Blk  \\\n",
       "count  733826.000000  733826.000000   733826.000000  733826.000000   \n",
       "mean      153.688580      20.551899       20.383095       8.590226   \n",
       "std        89.181256      15.385258       15.357587       4.769283   \n",
       "min         0.000000       0.000000        0.000000       0.000000   \n",
       "25%        71.000000       7.000000        7.000000       5.000000   \n",
       "50%       175.000000      17.000000       18.000000       9.000000   \n",
       "75%       231.000000      33.000000       33.000000      13.000000   \n",
       "max       310.000000      51.000000       51.000000      18.000000   \n",
       "\n",
       "         ArrTime_Blk  \n",
       "count  733826.000000  \n",
       "mean       10.215711  \n",
       "std         4.945489  \n",
       "min         0.000000  \n",
       "25%         6.000000  \n",
       "50%        11.000000  \n",
       "75%        14.000000  \n",
       "max        18.000000  \n",
       "\n",
       "[8 rows x 48 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Dimension reduced to:')\n",
    "print(len(data.columns))\n",
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All variables are now numeric. But above statistics makes sense only for originally numeric variables. For e.g. central tendency measures for Year, Airline ID etc. do not make much sense."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Machine Learning Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest is chosen for Machine Learning due to following reasons:\n",
    "(1) Medium data size    \n",
    "(2) Categorical nature of many variables       \n",
    "(3) Non-linearity of dataset (LinearRegression was attempted but had high MSE; Linear relation doesn't exist for many variables)      \n",
    "(4) Underlying complex dependencies between variables     \n",
    "(5) Non-parameteric approach without much assumptions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Creation of dataset for modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "47\n"
     ]
    }
   ],
   "source": [
    "# Data set up as predictors and target\n",
    "np.random.seed(10)\n",
    "Delay_YesNo = data['ArrDel15']\n",
    "data.drop(['ArrDel15'], axis=1, inplace=True)#Removing target variable\n",
    "print('Dimension reduced to:')\n",
    "print(len(data.columns))\n",
    "data_part2 = pd.DataFrame(data)\n",
    "data_part3 = pd.DataFrame(data)\n",
    "data_part4 = pd.DataFrame(data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PART 2: Model creation with redundant and collinear raw data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Train/test split\n",
    "Train:Test split = 80:20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(data_part2, Delay_YesNo, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Grid search\n",
    "Grid search is not performed for all the desired parameter combinations and values due to computational time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random forest Grid Search with non-redundant variables took [ 0:01:34.621699 ] seconds.\n"
     ]
    }
   ],
   "source": [
    "startTimeGS = datetime.now()\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "rf = RandomForestClassifier()\n",
    "param_grid = {\n",
    "                 'n_estimators': [10],\n",
    "#                 'min_samples_split': [2, 4],\n",
    "#                 'min_samples_leaf': [2, 4],\n",
    "#                 'max_features': ['sqrt', 'log2'],\n",
    "                 \"criterion\" : [\"gini\"]\n",
    "             }\n",
    "grid_rf = GridSearchCV(rf, param_grid, cv=3)\n",
    "grid_rf.fit(X_train, Y_train)\n",
    "bestModel = grid_rf.best_estimator_\n",
    "bestParameters = grid_rf.best_params_\n",
    "gridScores = grid_rf.grid_scores_\n",
    "print('Random forest Grid Search with non-redundant variables took [', datetime.now() - startTimeGS, '] seconds.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Best model, corresponding parameters and CV results\n",
    "Used for model selection in next step"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
      "            min_impurity_split=1e-07, min_samples_leaf=1,\n",
      "            min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
      "            n_estimators=10, n_jobs=1, oob_score=False, random_state=None,\n",
      "            verbose=0, warm_start=False)\n",
      "{'criterion': 'gini', 'n_estimators': 10}\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[mean: 0.93054, std: 0.00244, params: {'criterion': 'gini', 'n_estimators': 10}]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(bestModel)\n",
    "print(bestParameters)\n",
    "gridScores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Model selection\n",
    "- Best model and parameters from above are used to train the final model on entire training set\n",
    "- 3-fold Cross validation is performed to find the overall error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean cross validation score is: 0.930443224714\n",
      "Random forest training and testing with with non-redundant variables took [ 0:01:36.205567 ] seconds.\n"
     ]
    }
   ],
   "source": [
    "startTimeRF = datetime.now()\n",
    "rf = RandomForestClassifier(n_estimators = bestParameters.get('n_estimators'), \n",
    "#                            min_samples_split=bestParameters.get('min_samples_split'),\n",
    "#                           min_samples_leaf = bestParameters.get('min_samples_leaf'),\n",
    "#                           max_features = bestParameters.get('max_features'),\n",
    "                           criterion = bestParameters.get('criterion'))\n",
    "cv = cross_validation.KFold(len(X_train), n_folds=3, shuffle=True, random_state=2)\n",
    "cvScores = cross_val_score(rf, X_train, Y_train, cv=cv)\n",
    "print ('Mean cross validation score is: ' + str(np.mean(cvScores)))\n",
    "rf.fit(X_train, Y_train)\n",
    "print('Random forest training and testing with with non-redundant variables took [', \n",
    "      datetime.now() - startTimeRF, '] seconds.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Performance evaluation\n",
    "- Prediction is done on the hold out test set to evaluate performance\n",
    "- Confusion matrix, Accuracy and Recall are computed\n",
    "- ROC curve is also plotted to pictorically depict 'Area under curve (AUC)' as 'Accuracy' "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 93.79%\n",
      "Recall: 92.04%\n",
      "Confusion matrix:\n",
      "[[69940  3259]\n",
      " [ 5859 67708]]\n",
      "Area under the ROC curve: 0.0620820766617\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh4AAAGICAYAAAAOFOmJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nV8GuJMG02AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a007fd9f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAApUAAAGICAYAAAAQ8Rt/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0034547f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y_rf_pred = rf.predict(X_test)\n",
    "\n",
    "labels = [0, 1]\n",
    "cm = confusion_matrix(Y_test, Y_rf_pred,labels)\n",
    "\n",
    "print('Accuracy: ' + str(np.round(100*float(cm[0][0]+cm[1][1])/float((cm[0][0]+cm[1][1] + cm[1][0] + cm[0][1])),2))+'%')\n",
    "print('Recall: ' + str(np.round(100*float((cm[1][1]))/float((cm[1][0]+cm[1][1])),2))+'%')\n",
    "print('Confusion matrix:')\n",
    "print(cm)\n",
    "\n",
    "fpr, tpr, _ = roc_curve(Y_test, Y_rf_pred)\n",
    "auc = np.trapz(fpr,tpr)\n",
    "print('Area under the ROC curve: ' + str(auc))\n",
    "\n",
    "fig = plt.figure(1)\n",
    "plt.plot(fpr,tpr,color='green')\n",
    "plt.xlabel('False positive rate (FPR)')\n",
    "plt.ylabel('True positive rate (TPR)')\n",
    "plt.title('Receiver operating characteristic (ROC)')\n",
    "\n",
    "fig = plt.figure(2)\n",
    "ax = fig.add_subplot(111)\n",
    "cax = ax.matshow(cm)\n",
    "plt.title('Confusion matrix for Random Forest classifier with original data')\n",
    "fig.colorbar(cax)\n",
    "ax.set_xticklabels([''] + labels)\n",
    "ax.set_yticklabels([''] + labels)\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('True')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inferences 1:\n",
    "- High accuracy was obtained when model was first created.\n",
    "- This prompted cross validation with 3-fold in order to be convinced that it wasnt a chance performance (10-fold would have been better but not done due to computational time)\n",
    "- Accuracy improved when more data was added. More data is always better than any sophisticated model!\n",
    "- Inferences from exploratory analysis was critical in doing feature selection and obtaining good performance\n",
    "- E.g. Certain features (DepTimeBlk, Month etc.) with substantial variation were added to improve performance\n",
    "- Handling of class imbalance improved Accuracy and Recall respectively by more than 10% and 15%"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PART 3: Model creation with non-redundant raw data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Selection to remove redundant variables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Creation of data subset:\n",
    "Predictor and target variable columns along with some basic statistics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "47\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Year</th>\n",
       "      <th>Quarter</th>\n",
       "      <th>Month</th>\n",
       "      <th>DayofMonth</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>FlightDate</th>\n",
       "      <th>AirlineID</th>\n",
       "      <th>FlightNum</th>\n",
       "      <th>OriginAirportID</th>\n",
       "      <th>OriginAirportSeqID</th>\n",
       "      <th>...</th>\n",
       "      <th>Origin_Point</th>\n",
       "      <th>Origin_CityName</th>\n",
       "      <th>Origin_State</th>\n",
       "      <th>OriginState_Name</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Dest_CityName</th>\n",
       "      <th>Dest_State</th>\n",
       "      <th>Dest_StateName</th>\n",
       "      <th>DepTime_Blk</th>\n",
       "      <th>ArrTime_Blk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.00000</td>\n",
       "      <td>7.338260e+05</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>7.338260e+05</td>\n",
       "      <td>...</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2016.497052</td>\n",
       "      <td>2.476836</td>\n",
       "      <td>5.430509</td>\n",
       "      <td>15.794864</td>\n",
       "      <td>4.02110</td>\n",
       "      <td>2.016553e+07</td>\n",
       "      <td>19904.732889</td>\n",
       "      <td>2072.336574</td>\n",
       "      <td>12691.547570</td>\n",
       "      <td>1.269158e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>156.978973</td>\n",
       "      <td>150.878129</td>\n",
       "      <td>20.642058</td>\n",
       "      <td>20.475231</td>\n",
       "      <td>158.967048</td>\n",
       "      <td>153.688580</td>\n",
       "      <td>20.551899</td>\n",
       "      <td>20.383095</td>\n",
       "      <td>8.590226</td>\n",
       "      <td>10.215711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.499992</td>\n",
       "      <td>1.089987</td>\n",
       "      <td>3.269961</td>\n",
       "      <td>8.905223</td>\n",
       "      <td>2.03702</td>\n",
       "      <td>4.710840e+03</td>\n",
       "      <td>389.869122</td>\n",
       "      <td>1704.295317</td>\n",
       "      <td>1526.501453</td>\n",
       "      <td>1.526499e+05</td>\n",
       "      <td>...</td>\n",
       "      <td>86.031285</td>\n",
       "      <td>89.287731</td>\n",
       "      <td>15.351702</td>\n",
       "      <td>15.325205</td>\n",
       "      <td>86.132155</td>\n",
       "      <td>89.181256</td>\n",
       "      <td>15.385258</td>\n",
       "      <td>15.357587</td>\n",
       "      <td>4.769283</td>\n",
       "      <td>4.945489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>2.016070e+07</td>\n",
       "      <td>19393.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>10135.000000</td>\n",
       "      <td>1.013503e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>2.00000</td>\n",
       "      <td>2.016073e+07</td>\n",
       "      <td>19690.000000</td>\n",
       "      <td>706.000000</td>\n",
       "      <td>11292.000000</td>\n",
       "      <td>1.129202e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>70.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>71.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>6.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>4.00000</td>\n",
       "      <td>2.016103e+07</td>\n",
       "      <td>19805.000000</td>\n",
       "      <td>1613.000000</td>\n",
       "      <td>12892.000000</td>\n",
       "      <td>1.289204e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>170.000000</td>\n",
       "      <td>165.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>170.000000</td>\n",
       "      <td>175.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>11.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>6.00000</td>\n",
       "      <td>2.017013e+07</td>\n",
       "      <td>20304.000000</td>\n",
       "      <td>2878.000000</td>\n",
       "      <td>14027.000000</td>\n",
       "      <td>1.402702e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>229.000000</td>\n",
       "      <td>231.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>230.000000</td>\n",
       "      <td>231.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>14.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>7.00000</td>\n",
       "      <td>2.017043e+07</td>\n",
       "      <td>21171.000000</td>\n",
       "      <td>7439.000000</td>\n",
       "      <td>16218.000000</td>\n",
       "      <td>1.621801e+06</td>\n",
       "      <td>...</td>\n",
       "      <td>314.000000</td>\n",
       "      <td>310.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>314.000000</td>\n",
       "      <td>310.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>51.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 47 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Year        Quarter          Month     DayofMonth  \\\n",
       "count  733826.000000  733826.000000  733826.000000  733826.000000   \n",
       "mean     2016.497052       2.476836       5.430509      15.794864   \n",
       "std         0.499992       1.089987       3.269961       8.905223   \n",
       "min      2016.000000       1.000000       1.000000       1.000000   \n",
       "25%      2016.000000       2.000000       4.000000       8.000000   \n",
       "50%      2016.000000       3.000000       7.000000      16.000000   \n",
       "75%      2017.000000       3.000000       7.000000      24.000000   \n",
       "max      2017.000000       4.000000      10.000000      31.000000   \n",
       "\n",
       "          DayOfWeek    FlightDate      AirlineID      FlightNum  \\\n",
       "count  733826.00000  7.338260e+05  733826.000000  733826.000000   \n",
       "mean        4.02110  2.016553e+07   19904.732889    2072.336574   \n",
       "std         2.03702  4.710840e+03     389.869122    1704.295317   \n",
       "min         1.00000  2.016070e+07   19393.000000       1.000000   \n",
       "25%         2.00000  2.016073e+07   19690.000000     706.000000   \n",
       "50%         4.00000  2.016103e+07   19805.000000    1613.000000   \n",
       "75%         6.00000  2.017013e+07   20304.000000    2878.000000   \n",
       "max         7.00000  2.017043e+07   21171.000000    7439.000000   \n",
       "\n",
       "       OriginAirportID  OriginAirportSeqID      ...         Origin_Point  \\\n",
       "count    733826.000000        7.338260e+05      ...        733826.000000   \n",
       "mean      12691.547570        1.269158e+06      ...           156.978973   \n",
       "std        1526.501453        1.526499e+05      ...            86.031285   \n",
       "min       10135.000000        1.013503e+06      ...             0.000000   \n",
       "25%       11292.000000        1.129202e+06      ...            80.000000   \n",
       "50%       12892.000000        1.289204e+06      ...           170.000000   \n",
       "75%       14027.000000        1.402702e+06      ...           229.000000   \n",
       "max       16218.000000        1.621801e+06      ...           314.000000   \n",
       "\n",
       "       Origin_CityName   Origin_State  OriginState_Name    Destination  \\\n",
       "count    733826.000000  733826.000000     733826.000000  733826.000000   \n",
       "mean        150.878129      20.642058         20.475231     158.967048   \n",
       "std          89.287731      15.351702         15.325205      86.132155   \n",
       "min           0.000000       0.000000          0.000000       0.000000   \n",
       "25%          70.000000       7.000000          7.000000      80.000000   \n",
       "50%         165.000000      17.000000         18.000000     170.000000   \n",
       "75%         231.000000      33.000000         33.000000     230.000000   \n",
       "max         310.000000      51.000000         51.000000     314.000000   \n",
       "\n",
       "       Dest_CityName     Dest_State  Dest_StateName    DepTime_Blk  \\\n",
       "count  733826.000000  733826.000000   733826.000000  733826.000000   \n",
       "mean      153.688580      20.551899       20.383095       8.590226   \n",
       "std        89.181256      15.385258       15.357587       4.769283   \n",
       "min         0.000000       0.000000        0.000000       0.000000   \n",
       "25%        71.000000       7.000000        7.000000       5.000000   \n",
       "50%       175.000000      17.000000       18.000000       9.000000   \n",
       "75%       231.000000      33.000000       33.000000      13.000000   \n",
       "max       310.000000      51.000000       51.000000      18.000000   \n",
       "\n",
       "         ArrTime_Blk  \n",
       "count  733826.000000  \n",
       "mean       10.215711  \n",
       "std         4.945489  \n",
       "min         0.000000  \n",
       "25%         6.000000  \n",
       "50%        11.000000  \n",
       "75%        14.000000  \n",
       "max        18.000000  \n",
       "\n",
       "[8 rows x 47 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Data set up as predictors and target\n",
    "np.random.seed(10)\n",
    "rfDataOriginal = pd.DataFrame(data_part3)\n",
    "#Delay_YesNo = rfDataOriginal['ArrDel15']\n",
    "#rfDataOriginal.drop(['ArrDel15'], axis=1, inplace=True)#Removing target variable\n",
    "print('Dimension reduced to:')\n",
    "print(len(rfDataOriginal.columns))\n",
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Check for multi-collinearity of variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd0AAAFvCAYAAAAPPxEQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a002a8b400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(rfDataOriginal.corr())\n",
    "plt.title('Correlation matrix for MULTI-COLLINEAR data')\n",
    "corr = rfDataOriginal.corr()\n",
    "#corr"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: Darker cells indicate high collinearity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Year', 'FlightDate'),\n",
       " ('Quarter', 'Month'),\n",
       " ('OriginAirportID', 'OriginAirportSeqID'),\n",
       " ('OriginAirportID', 'Origin_Point'),\n",
       " ('OriginAirportSeqID', 'Origin_Point'),\n",
       " ('OriginStateFips', 'Origin_State'),\n",
       " ('OriginStateFips', 'OriginState_Name'),\n",
       " ('DestAirportID', 'DestAirportSeqID'),\n",
       " ('DestAirportID', 'Destination'),\n",
       " ('DestAirportSeqID', 'Destination'),\n",
       " ('DestStateFips', 'Dest_State'),\n",
       " ('DestStateFips', 'Dest_StateName'),\n",
       " ('CRSDepTime', 'DepTime'),\n",
       " ('CRSDepTime', 'WheelsOff'),\n",
       " ('CRSDepTime', 'DepTime_Blk'),\n",
       " ('DepTime', 'WheelsOff'),\n",
       " ('DepTime', 'DepTime_Blk'),\n",
       " ('WheelsOff', 'DepTime_Blk'),\n",
       " ('WheelsOn', 'ArrTime'),\n",
       " ('CRSArrTime', 'ArrTime_Blk'),\n",
       " ('CRSElapsedTime', 'ActualElapsedTime'),\n",
       " ('CRSElapsedTime', 'AirTime'),\n",
       " ('CRSElapsedTime', 'Distance'),\n",
       " ('CRSElapsedTime', 'DistanceGroup'),\n",
       " ('ActualElapsedTime', 'AirTime'),\n",
       " ('ActualElapsedTime', 'Distance'),\n",
       " ('ActualElapsedTime', 'DistanceGroup'),\n",
       " ('AirTime', 'Distance'),\n",
       " ('AirTime', 'DistanceGroup'),\n",
       " ('Distance', 'DistanceGroup'),\n",
       " ('Unique_Carrier', 'Carrier_Name'),\n",
       " ('Origin_State', 'OriginState_Name'),\n",
       " ('Dest_State', 'Dest_StateName')]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr = rfDataOriginal.corr()#Lists all pairs of highly collinear variables\n",
    "indices = np.where(corr > 0.8)\n",
    "indices = [(corr.columns[x], corr.columns[y]) for x, y in zip(*indices)\n",
    "                                        if x != y and x < y]\n",
    "indices"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Removal of only redundant columns\n",
    "Note: Collinearity is not the main consideration here(Part 3 deals with it)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Subsetting to remove redundant variables using results from above cell for quick filtering\n",
    "rfDataOriginal.drop(['OriginAirportSeqID','OriginCityMarketID', 'OriginStateFips', 'OriginWac', \n",
    "                     'DestAirportSeqID', 'DestCityMarketID', 'DestStateFips', 'DestWac',\n",
    "                     'CRSDepTime', 'CRSElapsedTime', 'CRSArrTime','Carrier_Name', \n",
    "                     'Origin_Point', 'Origin_CityName', 'Origin_State', 'OriginState_Name', \n",
    "                     'Destination', 'Dest_CityName', 'Dest_State', 'Dest_StateName'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension finally reduced to:\n",
      "27\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['Year',\n",
       " 'Quarter',\n",
       " 'Month',\n",
       " 'DayofMonth',\n",
       " 'DayOfWeek',\n",
       " 'FlightDate',\n",
       " 'AirlineID',\n",
       " 'FlightNum',\n",
       " 'OriginAirportID',\n",
       " 'DestAirportID',\n",
       " 'DepTime',\n",
       " 'TaxiOut',\n",
       " 'WheelsOff',\n",
       " 'WheelsOn',\n",
       " 'TaxiIn',\n",
       " 'ArrTime',\n",
       " 'Cancelled',\n",
       " 'Diverted',\n",
       " 'ActualElapsedTime',\n",
       " 'AirTime',\n",
       " 'Flights',\n",
       " 'Distance',\n",
       " 'DistanceGroup',\n",
       " 'Unique_Carrier',\n",
       " 'Tail_Number',\n",
       " 'DepTime_Blk',\n",
       " 'ArrTime_Blk']"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Dimension finally reduced to:')\n",
    "print(len(rfDataOriginal.columns))\n",
    "list(rfDataOriginal.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model building with non-redundant variables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Train/test split\n",
    "Train:Test split = 80:20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(rfDataOriginal, Delay_YesNo, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Grid search\n",
    "Grid search is not performed for all the desired parameter combinations and values due to computational time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random forest Grid Search with non-redundant variables took [ 0:01:19.389496 ] seconds.\n"
     ]
    }
   ],
   "source": [
    "startTimeGS = datetime.now()\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "rf = RandomForestClassifier()\n",
    "param_grid = {\n",
    "                 'n_estimators': [10],\n",
    "#                 'min_samples_split': [2, 4],\n",
    "#                 'min_samples_leaf': [2, 4],\n",
    "#                 'max_features': ['sqrt', 'log2'],\n",
    "                 \"criterion\" : [\"gini\"]\n",
    "             }\n",
    "grid_rf = GridSearchCV(rf, param_grid, cv=3)\n",
    "grid_rf.fit(X_train, Y_train)\n",
    "bestModel = grid_rf.best_estimator_\n",
    "bestParameters = grid_rf.best_params_\n",
    "gridScores = grid_rf.grid_scores_\n",
    "print('Random forest Grid Search with non-redundant variables took [', datetime.now() - startTimeGS, '] seconds.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Best model, corresponding parameters and CV results\n",
    "Used for model selection in next step"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
      "            min_impurity_split=1e-07, min_samples_leaf=1,\n",
      "            min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
      "            n_estimators=10, n_jobs=1, oob_score=False, random_state=None,\n",
      "            verbose=0, warm_start=False)\n",
      "{'criterion': 'gini', 'n_estimators': 10}\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[mean: 0.89131, std: 0.00067, params: {'criterion': 'gini', 'n_estimators': 10}]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(bestModel)\n",
    "print(bestParameters)\n",
    "gridScores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Model selection\n",
    "- Best model and parameters from above are used to train the final model on entire training set\n",
    "- 3-fold Cross validation is performed to find the overall error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean cross validation score is: 0.891683986507\n",
      "Random forest training and testing with with non-redundant variables took [ 0:01:20.739827 ] seconds.\n"
     ]
    }
   ],
   "source": [
    "startTimeRF = datetime.now()\n",
    "rf = RandomForestClassifier(n_estimators = bestParameters.get('n_estimators'), \n",
    "#                            min_samples_split=bestParameters.get('min_samples_split'),\n",
    "#                           min_samples_leaf = bestParameters.get('min_samples_leaf'),\n",
    "#                           max_features = bestParameters.get('max_features'),\n",
    "                           criterion = bestParameters.get('criterion'))\n",
    "cv = cross_validation.KFold(len(X_train), n_folds=3, shuffle=True, random_state=2)\n",
    "cvScores = cross_val_score(rf, X_train, Y_train, cv=cv)\n",
    "print ('Mean cross validation score is: ' + str(np.mean(cvScores)))\n",
    "rf.fit(X_train, Y_train)\n",
    "print('Random forest training and testing with with non-redundant variables took [', \n",
    "      datetime.now() - startTimeRF, '] seconds.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Performance evaluation\n",
    "- Prediction is done on the hold out test set to evaluate performance\n",
    "- Confusion matrix, Accuracy and Recall are computed\n",
    "- ROC curve is also plotted to pictorically depict 'Area under curve (AUC)' as 'Accuracy' "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 90.01%\n",
      "Recall: 87.14%\n",
      "Confusion matrix:\n",
      "[[67999  5200]\n",
      " [ 9460 64107]]\n",
      "Area under the ROC curve: 0.099814743252\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh4AAAGICAYAAAAOFOmJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n7wjpsQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a00441d160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAApUAAAGICAYAAAAQ8Rt/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0045a2f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y_rf_pred = rf.predict(X_test)\n",
    "\n",
    "labels = [0, 1]\n",
    "cm = confusion_matrix(Y_test, Y_rf_pred,labels)\n",
    "\n",
    "print('Accuracy: ' + str(np.round(100*float(cm[0][0]+cm[1][1])/float((cm[0][0]+cm[1][1] + cm[1][0] + cm[0][1])),2))+'%')\n",
    "print('Recall: ' + str(np.round(100*float((cm[1][1]))/float((cm[1][0]+cm[1][1])),2))+'%')\n",
    "print('Confusion matrix:')\n",
    "print(cm)\n",
    "\n",
    "fpr, tpr, _ = roc_curve(Y_test, Y_rf_pred)\n",
    "auc = np.trapz(fpr,tpr)\n",
    "print('Area under the ROC curve: ' + str(auc))\n",
    "\n",
    "fig = plt.figure(1)\n",
    "plt.plot(fpr,tpr,color='green')\n",
    "plt.xlabel('False positive rate (FPR)')\n",
    "plt.ylabel('True positive rate (TPR)')\n",
    "plt.title('Receiver operating characteristic (ROC)')\n",
    "\n",
    "fig = plt.figure(2)\n",
    "ax = fig.add_subplot(111)\n",
    "cax = ax.matshow(cm)\n",
    "plt.title('Confusion matrix for Random Forest classifier with original data')\n",
    "fig.colorbar(cax)\n",
    "ax.set_xticklabels([''] + labels)\n",
    "ax.set_yticklabels([''] + labels)\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('True')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inferences 2:\n",
    "- It appears that the removal of redundant variables has marginally reduced perfomrnace by 3-5%\n",
    "- This seems to indicate that Random Forest works well with highly-correlated data that has many interactions\n",
    "- In a nutshell, multicollinearity hasn't affected overall model prediction. But, it can cause overfitting and is not a concern here in the light of results from cross validation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PART 4: Model creation with PCA components"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Engineering by removing multi-collinear variables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Removal of redundant and highly collinear variables from dataset\n",
    "Highly collinear variables are removed before PCA as they amplify their correlation through the common underlying factor and distort PCA results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "27\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Year</th>\n",
       "      <th>Quarter</th>\n",
       "      <th>Month</th>\n",
       "      <th>DayofMonth</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>FlightDate</th>\n",
       "      <th>AirlineID</th>\n",
       "      <th>FlightNum</th>\n",
       "      <th>OriginAirportID</th>\n",
       "      <th>DestAirportID</th>\n",
       "      <th>...</th>\n",
       "      <th>Diverted</th>\n",
       "      <th>ActualElapsedTime</th>\n",
       "      <th>AirTime</th>\n",
       "      <th>Flights</th>\n",
       "      <th>Distance</th>\n",
       "      <th>DistanceGroup</th>\n",
       "      <th>Unique_Carrier</th>\n",
       "      <th>Tail_Number</th>\n",
       "      <th>DepTime_Blk</th>\n",
       "      <th>ArrTime_Blk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.00000</td>\n",
       "      <td>7.338260e+05</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>733826.0</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.0</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2016.497052</td>\n",
       "      <td>2.476836</td>\n",
       "      <td>5.430509</td>\n",
       "      <td>15.794864</td>\n",
       "      <td>4.02110</td>\n",
       "      <td>2.016553e+07</td>\n",
       "      <td>19904.732889</td>\n",
       "      <td>2072.336574</td>\n",
       "      <td>12691.547570</td>\n",
       "      <td>12727.301815</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>146.351322</td>\n",
       "      <td>119.988981</td>\n",
       "      <td>1.0</td>\n",
       "      <td>867.685536</td>\n",
       "      <td>3.939959</td>\n",
       "      <td>5.804210</td>\n",
       "      <td>2594.832654</td>\n",
       "      <td>8.590226</td>\n",
       "      <td>10.215711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.499992</td>\n",
       "      <td>1.089987</td>\n",
       "      <td>3.269961</td>\n",
       "      <td>8.905223</td>\n",
       "      <td>2.03702</td>\n",
       "      <td>4.710840e+03</td>\n",
       "      <td>389.869122</td>\n",
       "      <td>1704.295317</td>\n",
       "      <td>1526.501453</td>\n",
       "      <td>1529.135704</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>79.127785</td>\n",
       "      <td>76.084181</td>\n",
       "      <td>0.0</td>\n",
       "      <td>630.023799</td>\n",
       "      <td>2.479594</td>\n",
       "      <td>4.098074</td>\n",
       "      <td>1442.603771</td>\n",
       "      <td>4.769283</td>\n",
       "      <td>4.945489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>2.016070e+07</td>\n",
       "      <td>19393.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>10135.000000</td>\n",
       "      <td>10135.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>2.00000</td>\n",
       "      <td>2.016073e+07</td>\n",
       "      <td>19690.000000</td>\n",
       "      <td>706.000000</td>\n",
       "      <td>11292.000000</td>\n",
       "      <td>11292.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>87.000000</td>\n",
       "      <td>63.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>396.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1408.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>6.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>4.00000</td>\n",
       "      <td>2.016103e+07</td>\n",
       "      <td>19805.000000</td>\n",
       "      <td>1613.000000</td>\n",
       "      <td>12892.000000</td>\n",
       "      <td>12892.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>127.000000</td>\n",
       "      <td>101.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>692.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>2586.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>11.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>6.00000</td>\n",
       "      <td>2.017013e+07</td>\n",
       "      <td>20304.000000</td>\n",
       "      <td>2878.000000</td>\n",
       "      <td>14027.000000</td>\n",
       "      <td>14057.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>181.000000</td>\n",
       "      <td>152.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1119.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>3895.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>14.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>7.00000</td>\n",
       "      <td>2.017043e+07</td>\n",
       "      <td>21171.000000</td>\n",
       "      <td>7439.000000</td>\n",
       "      <td>16218.000000</td>\n",
       "      <td>16218.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>744.000000</td>\n",
       "      <td>704.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4983.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>4999.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Year        Quarter          Month     DayofMonth  \\\n",
       "count  733826.000000  733826.000000  733826.000000  733826.000000   \n",
       "mean     2016.497052       2.476836       5.430509      15.794864   \n",
       "std         0.499992       1.089987       3.269961       8.905223   \n",
       "min      2016.000000       1.000000       1.000000       1.000000   \n",
       "25%      2016.000000       2.000000       4.000000       8.000000   \n",
       "50%      2016.000000       3.000000       7.000000      16.000000   \n",
       "75%      2017.000000       3.000000       7.000000      24.000000   \n",
       "max      2017.000000       4.000000      10.000000      31.000000   \n",
       "\n",
       "          DayOfWeek    FlightDate      AirlineID      FlightNum  \\\n",
       "count  733826.00000  7.338260e+05  733826.000000  733826.000000   \n",
       "mean        4.02110  2.016553e+07   19904.732889    2072.336574   \n",
       "std         2.03702  4.710840e+03     389.869122    1704.295317   \n",
       "min         1.00000  2.016070e+07   19393.000000       1.000000   \n",
       "25%         2.00000  2.016073e+07   19690.000000     706.000000   \n",
       "50%         4.00000  2.016103e+07   19805.000000    1613.000000   \n",
       "75%         6.00000  2.017013e+07   20304.000000    2878.000000   \n",
       "max         7.00000  2.017043e+07   21171.000000    7439.000000   \n",
       "\n",
       "       OriginAirportID  DestAirportID      ...        Diverted  \\\n",
       "count    733826.000000  733826.000000      ...        733826.0   \n",
       "mean      12691.547570   12727.301815      ...             0.0   \n",
       "std        1526.501453    1529.135704      ...             0.0   \n",
       "min       10135.000000   10135.000000      ...             0.0   \n",
       "25%       11292.000000   11292.000000      ...             0.0   \n",
       "50%       12892.000000   12892.000000      ...             0.0   \n",
       "75%       14027.000000   14057.000000      ...             0.0   \n",
       "max       16218.000000   16218.000000      ...             0.0   \n",
       "\n",
       "       ActualElapsedTime        AirTime   Flights       Distance  \\\n",
       "count      733826.000000  733826.000000  733826.0  733826.000000   \n",
       "mean          146.351322     119.988981       1.0     867.685536   \n",
       "std            79.127785      76.084181       0.0     630.023799   \n",
       "min            15.000000       7.000000       1.0      31.000000   \n",
       "25%            87.000000      63.000000       1.0     396.000000   \n",
       "50%           127.000000     101.000000       1.0     692.000000   \n",
       "75%           181.000000     152.000000       1.0    1119.000000   \n",
       "max           744.000000     704.000000       1.0    4983.000000   \n",
       "\n",
       "       DistanceGroup  Unique_Carrier    Tail_Number    DepTime_Blk  \\\n",
       "count  733826.000000   733826.000000  733826.000000  733826.000000   \n",
       "mean        3.939959        5.804210    2594.832654       8.590226   \n",
       "std         2.479594        4.098074    1442.603771       4.769283   \n",
       "min         1.000000        0.000000       0.000000       0.000000   \n",
       "25%         2.000000        2.000000    1408.000000       5.000000   \n",
       "50%         3.000000        5.000000    2586.000000       9.000000   \n",
       "75%         5.000000       10.000000    3895.000000      13.000000   \n",
       "max        11.000000       11.000000    4999.000000      18.000000   \n",
       "\n",
       "         ArrTime_Blk  \n",
       "count  733826.000000  \n",
       "mean       10.215711  \n",
       "std         4.945489  \n",
       "min         0.000000  \n",
       "25%         6.000000  \n",
       "50%        11.000000  \n",
       "75%        14.000000  \n",
       "max        18.000000  \n",
       "\n",
       "[8 rows x 27 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(10)\n",
    "rfDataPCA = pd.DataFrame(data_part4)\n",
    "#Subsetting based on earlier feature selection of variables (derived from correlation)\n",
    "rfDataPCA.drop(['OriginAirportSeqID','OriginCityMarketID', 'OriginStateFips', 'OriginWac', \n",
    "                     'DestAirportSeqID', 'DestCityMarketID', 'DestStateFips', 'DestWac',\n",
    "                     'CRSDepTime', 'CRSElapsedTime', 'CRSArrTime','Carrier_Name', \n",
    "                     'Origin_Point', 'Origin_CityName', 'Origin_State', 'OriginState_Name', \n",
    "                     'Destination', 'Dest_CityName', 'Dest_State', 'Dest_StateName'], axis=1, inplace=True)\n",
    "print('Dimension reduced to:')\n",
    "print(len(rfDataPCA.columns))\n",
    "list(rfDataPCA.columns)\n",
    "rfDataPCA.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Year', 'FlightDate'),\n",
       " ('Quarter', 'Month'),\n",
       " ('OriginAirportID', 'OriginAirportSeqID'),\n",
       " ('OriginAirportID', 'Origin_Point'),\n",
       " ('OriginAirportID', 'Origin_CityName'),\n",
       " ('OriginAirportSeqID', 'Origin_Point'),\n",
       " ('OriginAirportSeqID', 'Origin_CityName'),\n",
       " ('OriginStateFips', 'Origin_State'),\n",
       " ('OriginStateFips', 'OriginState_Name'),\n",
       " ('DestAirportID', 'DestAirportSeqID'),\n",
       " ('DestAirportID', 'Destination'),\n",
       " ('DestAirportID', 'Dest_CityName'),\n",
       " ('DestAirportSeqID', 'Destination'),\n",
       " ('DestAirportSeqID', 'Dest_CityName'),\n",
       " ('DestStateFips', 'Dest_State'),\n",
       " ('DestStateFips', 'Dest_StateName'),\n",
       " ('CRSDepTime', 'DepTime'),\n",
       " ('CRSDepTime', 'WheelsOff'),\n",
       " ('CRSDepTime', 'DepTime_Blk'),\n",
       " ('CRSDepTime', 'ArrTime_Blk'),\n",
       " ('DepTime', 'WheelsOff'),\n",
       " ('DepTime', 'DepTime_Blk'),\n",
       " ('DepTime', 'ArrTime_Blk'),\n",
       " ('WheelsOff', 'DepTime_Blk'),\n",
       " ('WheelsOff', 'ArrTime_Blk'),\n",
       " ('WheelsOn', 'CRSArrTime'),\n",
       " ('WheelsOn', 'ArrTime'),\n",
       " ('WheelsOn', 'ArrTime_Blk'),\n",
       " ('CRSArrTime', 'ArrTime_Blk'),\n",
       " ('CRSElapsedTime', 'ActualElapsedTime'),\n",
       " ('CRSElapsedTime', 'AirTime'),\n",
       " ('CRSElapsedTime', 'Distance'),\n",
       " ('CRSElapsedTime', 'DistanceGroup'),\n",
       " ('ActualElapsedTime', 'AirTime'),\n",
       " ('ActualElapsedTime', 'Distance'),\n",
       " ('ActualElapsedTime', 'DistanceGroup'),\n",
       " ('AirTime', 'Distance'),\n",
       " ('AirTime', 'DistanceGroup'),\n",
       " ('Distance', 'DistanceGroup'),\n",
       " ('Unique_Carrier', 'Carrier_Name'),\n",
       " ('Origin_Point', 'Origin_CityName'),\n",
       " ('Origin_State', 'OriginState_Name'),\n",
       " ('Destination', 'Dest_CityName'),\n",
       " ('Dest_State', 'Dest_StateName'),\n",
       " ('DepTime_Blk', 'ArrTime_Blk')]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indices = np.where(corr > 0.7)\n",
    "indices = [(corr.columns[x], corr.columns[y]) for x, y in zip(*indices)\n",
    "                                        if x != y and x < y]\n",
    "indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rfDataPCA.drop(['FlightDate','DepTime','DepTime_Blk','ActualElapsedTime','AirTime','DistanceGroup','DepTime_Blk'], \n",
    "          axis=1, inplace=True)#Further reduction of multicollinearity for PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAc4AAAFvCAYAAADDpiv7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJE4AABKQOAEASPD/AVwBN9Ow+U/oAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0312c2668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Check for collinearity\n",
    "plt.matshow(rfDataPCA.corr())\n",
    "plt.title('Correlation matrix for NON-COLLINEAR data')\n",
    "corr = rfDataPCA.corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: Darker cells indicate high collinearity. It is seen here only between the variable and itself."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Quarter', 'Month'),\n",
       " ('WheelsOff', 'ArrTime_Blk'),\n",
       " ('WheelsOn', 'ArrTime'),\n",
       " ('WheelsOn', 'ArrTime_Blk')]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indices = np.where(corr > 0.7)\n",
    "indices = [(corr.columns[x], corr.columns[y]) for x, y in zip(*indices)\n",
    "                                        if x != y and x < y]\n",
    "indices"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Data ready for PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "21\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Year</th>\n",
       "      <th>Quarter</th>\n",
       "      <th>Month</th>\n",
       "      <th>DayofMonth</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>AirlineID</th>\n",
       "      <th>FlightNum</th>\n",
       "      <th>OriginAirportID</th>\n",
       "      <th>DestAirportID</th>\n",
       "      <th>TaxiOut</th>\n",
       "      <th>...</th>\n",
       "      <th>WheelsOn</th>\n",
       "      <th>TaxiIn</th>\n",
       "      <th>ArrTime</th>\n",
       "      <th>Cancelled</th>\n",
       "      <th>Diverted</th>\n",
       "      <th>Flights</th>\n",
       "      <th>Distance</th>\n",
       "      <th>Unique_Carrier</th>\n",
       "      <th>Tail_Number</th>\n",
       "      <th>ArrTime_Blk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.00000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.0</td>\n",
       "      <td>733826.0</td>\n",
       "      <td>733826.0</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "      <td>733826.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2016.497052</td>\n",
       "      <td>2.476836</td>\n",
       "      <td>5.430509</td>\n",
       "      <td>15.794864</td>\n",
       "      <td>4.02110</td>\n",
       "      <td>19904.732889</td>\n",
       "      <td>2072.336574</td>\n",
       "      <td>12691.547570</td>\n",
       "      <td>12727.301815</td>\n",
       "      <td>18.214939</td>\n",
       "      <td>...</td>\n",
       "      <td>1503.319324</td>\n",
       "      <td>8.147401</td>\n",
       "      <td>1503.626556</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>867.685536</td>\n",
       "      <td>5.804210</td>\n",
       "      <td>2594.832654</td>\n",
       "      <td>10.215711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.499992</td>\n",
       "      <td>1.089987</td>\n",
       "      <td>3.269961</td>\n",
       "      <td>8.905223</td>\n",
       "      <td>2.03702</td>\n",
       "      <td>389.869122</td>\n",
       "      <td>1704.295317</td>\n",
       "      <td>1526.501453</td>\n",
       "      <td>1529.135704</td>\n",
       "      <td>11.933518</td>\n",
       "      <td>...</td>\n",
       "      <td>570.385123</td>\n",
       "      <td>7.499421</td>\n",
       "      <td>578.141924</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>630.023799</td>\n",
       "      <td>4.098074</td>\n",
       "      <td>1442.603771</td>\n",
       "      <td>4.945489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>19393.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>10135.000000</td>\n",
       "      <td>10135.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>2.00000</td>\n",
       "      <td>19690.000000</td>\n",
       "      <td>706.000000</td>\n",
       "      <td>11292.000000</td>\n",
       "      <td>11292.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1117.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>1119.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>396.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1408.000000</td>\n",
       "      <td>6.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2016.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>4.00000</td>\n",
       "      <td>19805.000000</td>\n",
       "      <td>1613.000000</td>\n",
       "      <td>12892.000000</td>\n",
       "      <td>12892.000000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1559.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>1600.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>692.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>2586.000000</td>\n",
       "      <td>11.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>6.00000</td>\n",
       "      <td>20304.000000</td>\n",
       "      <td>2878.000000</td>\n",
       "      <td>14027.000000</td>\n",
       "      <td>14057.000000</td>\n",
       "      <td>21.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1954.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>1957.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1119.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>3895.000000</td>\n",
       "      <td>14.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2017.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>7.00000</td>\n",
       "      <td>21171.000000</td>\n",
       "      <td>7439.000000</td>\n",
       "      <td>16218.000000</td>\n",
       "      <td>16218.000000</td>\n",
       "      <td>176.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2400.000000</td>\n",
       "      <td>413.000000</td>\n",
       "      <td>2400.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4983.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>4999.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Year        Quarter          Month     DayofMonth  \\\n",
       "count  733826.000000  733826.000000  733826.000000  733826.000000   \n",
       "mean     2016.497052       2.476836       5.430509      15.794864   \n",
       "std         0.499992       1.089987       3.269961       8.905223   \n",
       "min      2016.000000       1.000000       1.000000       1.000000   \n",
       "25%      2016.000000       2.000000       4.000000       8.000000   \n",
       "50%      2016.000000       3.000000       7.000000      16.000000   \n",
       "75%      2017.000000       3.000000       7.000000      24.000000   \n",
       "max      2017.000000       4.000000      10.000000      31.000000   \n",
       "\n",
       "          DayOfWeek      AirlineID      FlightNum  OriginAirportID  \\\n",
       "count  733826.00000  733826.000000  733826.000000    733826.000000   \n",
       "mean        4.02110   19904.732889    2072.336574     12691.547570   \n",
       "std         2.03702     389.869122    1704.295317      1526.501453   \n",
       "min         1.00000   19393.000000       1.000000     10135.000000   \n",
       "25%         2.00000   19690.000000     706.000000     11292.000000   \n",
       "50%         4.00000   19805.000000    1613.000000     12892.000000   \n",
       "75%         6.00000   20304.000000    2878.000000     14027.000000   \n",
       "max         7.00000   21171.000000    7439.000000     16218.000000   \n",
       "\n",
       "       DestAirportID        TaxiOut      ...             WheelsOn  \\\n",
       "count  733826.000000  733826.000000      ...        733826.000000   \n",
       "mean    12727.301815      18.214939      ...          1503.319324   \n",
       "std      1529.135704      11.933518      ...           570.385123   \n",
       "min     10135.000000       1.000000      ...             1.000000   \n",
       "25%     11292.000000      11.000000      ...          1117.000000   \n",
       "50%     12892.000000      15.000000      ...          1559.000000   \n",
       "75%     14057.000000      21.000000      ...          1954.000000   \n",
       "max     16218.000000     176.000000      ...          2400.000000   \n",
       "\n",
       "              TaxiIn        ArrTime  Cancelled  Diverted   Flights  \\\n",
       "count  733826.000000  733826.000000   733826.0  733826.0  733826.0   \n",
       "mean        8.147401    1503.626556        0.0       0.0       1.0   \n",
       "std         7.499421     578.141924        0.0       0.0       0.0   \n",
       "min         1.000000       1.000000        0.0       0.0       1.0   \n",
       "25%         5.000000    1119.000000        0.0       0.0       1.0   \n",
       "50%         6.000000    1600.000000        0.0       0.0       1.0   \n",
       "75%         9.000000    1957.000000        0.0       0.0       1.0   \n",
       "max       413.000000    2400.000000        0.0       0.0       1.0   \n",
       "\n",
       "            Distance  Unique_Carrier    Tail_Number    ArrTime_Blk  \n",
       "count  733826.000000   733826.000000  733826.000000  733826.000000  \n",
       "mean      867.685536        5.804210    2594.832654      10.215711  \n",
       "std       630.023799        4.098074    1442.603771       4.945489  \n",
       "min        31.000000        0.000000       0.000000       0.000000  \n",
       "25%       396.000000        2.000000    1408.000000       6.000000  \n",
       "50%       692.000000        5.000000    2586.000000      11.000000  \n",
       "75%      1119.000000       10.000000    3895.000000      14.000000  \n",
       "max      4983.000000       11.000000    4999.000000      18.000000  \n",
       "\n",
       "[8 rows x 21 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Dimension reduced to:')\n",
    "print(len(rfDataPCA.columns))\n",
    "rfDataPCA.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dimension reduction with PCA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This was attempted to check if classification performance would be better upon reducing dimensions with PCA. My initial thought was that PCA, an unsupervised approach, would help supervised learning by producing a simpler hypothesis space and reduce overfitting. It was also assumed that it would reduce noise in data and computational complexity. As a pre-requisite to PCA, collinearity was removed assuming it would give better model performance.\n",
    "\n",
    "PCA is done on three variants of the dataset to identify which one needs to be used for modeling:\n",
    "- PCA 1: Dataset with Multi-Collinear variables: Highly collinear variables have amplified contribution in PCA\n",
    "- PCA 2: Dataset with Redundant variables removed (i.e. repeated ID, dep time, arr time etc.)\n",
    "- PCA 3: Dataset without any Non-Collinear variables\n",
    "\n",
    "Due to limitation in training time, results are demonstrated only for model creation with PCs derived from non-collinear data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### PCA 1: On HIGHLY MULTI-COLLINEAR data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "47\n"
     ]
    }
   ],
   "source": [
    "#Data set up as predictors and target\n",
    "dataOriginal = pd.DataFrame(data_part4)\n",
    "#Delay_YesNo = dataOriginal['ArrDel15']\n",
    "#dataOriginal.drop(['ArrDel15'], axis=1, inplace=True)#Removing target variable\n",
    "print('Dimension reduced to:')\n",
    "print(len(dataOriginal.columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[  5.12101357e-01   4.87226232e-01   4.75020513e-04   6.33555584e-05\n",
      "   4.42966325e-05   2.64195699e-05   2.23766071e-05   1.99034512e-05]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ 51.21,  99.93,  99.98,  99.99,  99.99,  99.99,  99.99,  99.99])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca = PCA(n_components=8)\n",
    "dataOriginal_reduced = pca.fit_transform(dataOriginal)\n",
    "print(pca.explained_variance_ratio_)\n",
    "np.cumsum(np.round(pca.explained_variance_ratio_, decimals=4)*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1a004ae3fd0>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABlUAAAHVCAYAAABoh311AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxqIKcWFaVwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a00440e630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(1, figsize=(20, 5))\n",
    "plt.plot(pca.explained_variance_, linewidth=3)\n",
    "plt.axis('tight')\n",
    "plt.title('Scree plot for PCA with COLLINEAR data')\n",
    "plt.xlabel('No. of principal components')\n",
    "plt.ylabel('Explained variance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: Elbow is used to choose the number of principal components (Here, #PCs chosen = 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### PCA 2: On PARTIALLY MULTI-COLLINEAR data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "27\n"
     ]
    }
   ],
   "source": [
    "#Data set up as predictors and target\n",
    "dataPartialyCollinear = pd.DataFrame(rfDataOriginal)\n",
    "print('Dimension reduced to:')\n",
    "print(len(dataPartialyCollinear.columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.66064717  0.08795286  0.07384746  0.0677889   0.05942026  0.02682255\n",
      "  0.01072468  0.0077272 ]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ 66.06,  74.86,  82.24,  89.02,  94.96,  97.64,  98.71,  99.48])"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca = PCA(n_components=8)\n",
    "dataPartialyCollinear_reduced = pca.fit_transform(dataPartialyCollinear)\n",
    "print(pca.explained_variance_ratio_)\n",
    "np.cumsum(np.round(pca.explained_variance_ratio_, decimals=4)*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1a004aa87b8>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABlUAAAHVCAYAAABoh311AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a004b30128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(1, figsize=(20, 5))\n",
    "plt.plot(pca.explained_variance_, linewidth=3)\n",
    "plt.axis('tight')\n",
    "plt.title('Scree plot for PCA with FEATURE SELECTED data')\n",
    "plt.xlabel('No. of principal components')\n",
    "plt.ylabel('Explained variance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: Elbow is used to choose the number of principal components (Here, #PCs chosen = 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### PCA 3: On LOW MULTI-COLLINEAR data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimension reduced to:\n",
      "21\n"
     ]
    }
   ],
   "source": [
    "#Data set up as predictors and target\n",
    "dataNotCollinear = pd.DataFrame(rfDataPCA)\n",
    "#Delay_YesNo = dataNotCollinear['ArrDel15']\n",
    "#dataNotCollinear.drop(['ArrDel15'], axis=1, inplace=True)#Removing target variable\n",
    "print('Dimension reduced to:')\n",
    "print(len(dataNotCollinear.columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.26567194  0.22292369  0.20453226  0.17938456  0.06765603  0.03145279\n",
      "  0.01417245  0.0128247 ]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ 26.57,  48.86,  69.31,  87.25,  94.02,  97.17,  98.59,  99.87])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca = PCA(n_components=8)\n",
    "dataNotCollinear_reduced = pca.fit_transform(dataNotCollinear)\n",
    "print(pca.explained_variance_ratio_)\n",
    "np.cumsum(np.round(pca.explained_variance_ratio_, decimals=4)*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1a004422128>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABnUAAAHVCAYAAAAn2n6lAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzs3Xmc1VX9+PHXm0VRUFAQUZTNXcAMKvPrLpqmuaRlUZqW\nmZZpmlY/zXLLtCzNMktTW8wttzS3LHHLXdES92RTVEBERREQOL8/zuc2l5l7Z+bOMDMMvJ6Px3nc\n4XM+5/N538+9d0Y/73veJ1JKSJIkSZIkSZIkadnWpaMDkCRJkiRJkiRJUtNM6kiSJEmSJEmSJHUC\nJnUkSZIkSZIkSZI6AZM6kiRJkiRJkiRJnYBJHUmSJEmSJEmSpE7ApI4kSZIkSZIkSVInYFJHkiRJ\nkiRJkiSpEzCpI0mSJEmSJEmS1AmY1JEkSZIkSZIkSeoETOpIkiSpzUTE4IhYHBGXdnAcR0fE0xEx\nt4jn6I6MZ0USEX8orvmgGsbsUIz5YVvGJkm1iohXIuKFjo5DkiStuEzqSJIk1SgiukTEYRFxd0TM\niogFETE9Iv4dEb+LiL06OsblTUQcXNzk/1ILxn4e+AXwPnAucArw0NKNsFlxTC6eQ6ktiojZEXF/\nRHwjIrpWGbdeRJwVEY9FxJtl77d/FMmq1Rs55xfLzrdL2z27RqWilcfVrsm+ete9UvtS2b4nN2P/\ncY2cq9FrHhFjmnH8+u+TdSNig+LfVW8mR0TXYp8FzTjnguLm9LURsU2V4x3ajPiadXO7yrHmFZ+L\nP0XEiCrjTm9GDHc0J4bieH9uxvEuqnBNF0fESxHRvcpxX4mIRY2cd+OI+EVEjC/7HM+KiIci4qcR\nsWWFMT8qzntiI9fkq1XOV7reP2xkbLOuZ/HcmhrzhbL92+39FhH/qtf/QXF9n42IqyLiSxGxarXX\npYVS07s0rux91ez3riRJUkm3jg5AkiSpM4mILsAtwG7A7OLnV4CVgOHAWGAT4G8dFeNyrKU30vYs\nxu6ZUpq+FOOpVSm5cR7wFtAVGArsD2wN7Ax8pnxAccP2V+T317+BK8jvuzWB/yMnq04C+lc552HA\nYiCArwH/XJpPqJn+H3AmMK0Dzl0ukRN6UaHvyQrb7i5aJZMbOU9T13xiEUe5NYGjgTfJ74/yGBPw\nDrBKI+dsronAn4qfewIfAfYD9o2Iz6SU/lpl3Hjgpip9b9YYQ/mxegPbAgcC+0fETimlRyqMScBd\nwL1VjjmxhvOXPoc3AP9pJMZK44YARwHnVOmvKCJOA04kv66PU/c5Xh3YojjmcRFxRErpdxVibey5\nnBoRV6SU5tYQU63XsxTHueT3YiWVrmV7vN9Ksf0emEq+xqsBGwC7AAcAZ0bEl1NKJlAkSdJywaSO\nJElSbcaSEzpPADuklN4t74yIHsBWHRHYcq7SjfjmWheggxM65X6RUppa+kdEnAk8Bnw6IrZLKd1X\nbP8icBH5JuanU0q31z9QRHwMuKDSSSJiY2B74B/kpMHeEbFWSmnm0n5CjSmue/1r35rXszWxnF7D\n7nenlE6r5fjNueYppUnAafXGbUCR1KkWY0SsXUssVUys/5wi4vvA6cBPgKo32Wu9Fo1ocKyI+B1w\naBHHblXGjUsp/XgpxQBwfUrpihr2n0VOxH4/Ii5NKb3VnEERcTrwfWASMLZS0ioi1gKOBfrUEA/A\nf4ENge8Ap1Y6fSNjW3I9z0kpvVrD/u35frs0pfRAvXOtTN21+WtEjEkpPVjjcdtKh/wOlCRJywfL\nr0mSJNXm/8jfCv5j/YQOQEppXkrpnkoDI+JzEXFnUXLn/YiYFBFXRMTosn3+V2YsInaPiLsi4q36\nZX0iYpPIa5VMjYj5EfF6RFxe3FSudO5VIuKEiHgiIt6NiDkR8UDk0mTNFrlU0sSIWD0izi/K6bwf\neb2ao2o81oCI+HVxHeZHxIyIuC4iRtXb7y6gVKartD5LqSxV1XVaoiijBeyU/1k3rt5+YyLi9uJ1\nmRcRz0fEmVGhrFnkknuLIqJ7RPwwIp4rxrS4jFhK6VnqZoR8rDhPL+CX5Pfa5yoldIqxj5Bn+VTy\nNeq+wf4H8myfQ2qJLSJejYiXK2yfUlzL79fbvnux/ZSybUusqRMRJ5O/wZ+AQ+qVTmpQXi8iPhQR\nt0QuVfde8RpUe84drdXXvANcUjxuWOk9384xfLSDzt8c7wJnAGsAP2jOgIjYkDxTbR7wySqzkEgp\nzUwpnQj8vIZ4SrP+ppNn+QyoYWxHarf3W0ppfkrpR+SZgj3I16vZom4ttnnF37rzImK1Kvv2jojv\nRsS4Yt/5kctk3lAk38v3PRRYQH4Nd6n3O/DEsv2+UvxNfCnyenBvRcR9ETG21mshSZKWL87UkSRJ\nqs0s8jdsKyZPqomIPwBfAmYC1xWP65ETDs+RS/KUJOCzwO7ArcBvgEFlx9q9OEY3cpm3/xbH2g/Y\nMyJ2TCk9WbZ/b3KpnQ+Ry9pcQv5yz27AFRGxeUqpuQvSJ/KN6n+SSyddWfx7f+C8iNg4pdRkcici\nhgD3AwOAceRyROsXz3vPiNgvpXRrsfvvyaWK9iF/s7v03BK5jFk1dxX7fJl8/U4hv3b/K0kUEYeT\nZ7q8C1wDzAB2BL4HfCoitkkplZcbKo29jlxK6DZyGacZTT3nJpS+tV06/mfJN48fTCnd2djAlNIH\nDQ6W1/34EvB2Ed+q5BvGXwXOriGuO4EvFK/rC8WxNyC/VgkYQ77RXbJLsb285Fj9ElJ3kd87x5Bf\ny/Jv69cvg/ZR8mvxAPA78uv4GeCfEbFlSunFGp5Lm1qK17wjLezg8zd4Ly9jfgV8E/hGRJxfzLpq\nzKHk2T2XlT4/jUkpLa4xnneBH5Jn9J1OLv3XmbTX++2nwHHA6IjYqDm/NyLi18DXyWUjf0t+b+5L\nTrx3J6/RVm4EeQbePeTScW8Bg4G9gT0i4pMppdJaXI8X+/6QJUvUwZIl8X5LnhV8N/A60A/YA7g8\nIjasceahJElajpjUkSRJqs315JvMXy++ZXwD8Hh5Oa36IuJr5Ju9DwO7ls/wiYig4XooAXyS/M3u\nf9Q7Vh9yIuVdYPuU0vNlfZsX57iYnHAoOY+c0PluSunnZfuvBNwInBgR16aUqq0vUd86wEvANqWE\nQjH74jHyzc6rU0r/auIYF5ITOt9PKZ1VFtMFwH3AHyNicEppbkrpT8V12gf4a0rpT5UPuaSU0r3A\nvRGxEzCo/g2wYubIecAc4KPlN/rKbuj9FDii3qGDnFwYnlKa3ZxYGhMRw4Edin8+XDxuQ06ENJrQ\nacT+5BuAv00pzQfmR8TfgP0ir1tyVzOPMw74Ijl5U7opPaZ4/AewQ0T0SCnNK7btTL7Z+VC1A6aU\n7o2IKRRJnSbKLO0BHJJSuqy0ofg8/Rb4FvkGe7MV79P6JqeU/lhh+07F+66S21NKD9fbtrSueXsr\nvb+fqLIuC8CoKtcO4IH6v6da4PDi8b4q/QGMKRJnldyaUnq0xnPuHxEbVem7PKX03/obU0ofRMQJ\n5N/BZwJNzXQszexsy9f+UvJn6ZCI+EVK6elmjGnp9TwuIqqtqXNGSqk5SZp2f7+llN6JiCfIpVE/\nBjSa1ImI7ci//58Htiol9iPiJHLSpT8N1xZ6Clin/t+EiFgfeIS8HtGHiniejIinKJI6jfwO3KR+\n4rB4ze4glwG8MKXU2i8USJKkTsikjiRJUg2KmzFfJCcDvkhe4JuIeJN8s+fSlNLN9YYdRb6xd3j9\nkm0ppUTD9UYgJy8q3bg6mLy49jfLEzrFsZ6JvDbFtyJi05TScxGxZhHnY+UJnWL/BRHxPfKMnS9Q\nfdHwSk4onyGSUnor8toRvyfPjKma1ImIgcCuwBTqzWBIKT0UEVcWMe8H/LmGmGp1EPkb12dX+Ob2\n98mv7UERcVS92TAJOKkVCZ1jI+It8jf4hwGfJpcGuj6ldH+xzzrF4ystPMdhRZzlyYo/kBMPX6P5\nN5lLSaUx5BljpZ9fJ5eH+wR5oft/Fu+1DwF3NPPmbnP8qzyhU7gUOJ+iVF2NKs1Iu4clr1PJ9kWr\nZDZ1CbiSpXXN29KwspvlPckzoXYgr9tUP3lZ7sNFq+Tn5ARfc5XfsO8NbAeMJi9y/91Gxu1YtEpm\nArUkdYI862LfKv2PkmdANpBSujoijgU+GxHnVCupViiVRJvWIICIoeTf5+Wz2N5MKf2qqeDrxbM4\nIr4L3Ez+fbpHM4fuSO3X85hGjnc2DWfeLAvvt5LSa7BWM/b9Cvl1Ob18pmZKaX5RHq3B+evN6Czf\n/nJEXA8cEREDUkqvNzfgSjPBisTiBeTfTTsDVzX3eJIkaflhUkeSJKlGKaVrI+IGcum0bck3n7Yl\nzyTZNyL+mFL6MkBErAoMB16vYSYMVL9B+fHiccsq32QulYXbjFzW7aPk5EGqsv9KZfs310Kg0mLT\ndxeP1W7GUa//vpTSogr948gJlQ/TtkmdUhwNbrYXSaonyDecNyV/C7tcrbMCyh1dOg15xtW/gcvI\ns5darSiPtiPwXL3ZJLeTkzH7RsSaKaU3mzpWSmlqRExkyZu/O5K/KX4v+b0whlxubWfyzfKlmbx4\nvP6GlNLCiJhOLk9Xk5RS1xp2P6W55Y2W5jVvY0NpmNiaBeyUUprQyLhLUkpfW0oxVLphPwnYLqX0\napUxpUTqmU0dPCK+DdRf9+S6es8vAQemlK5sZsz1HU9+//+c/DuiJYaRX4tSUifIiaSakjoAKaVb\nI2IcsFtEjGmqZCM1XM961kspvVbD/svC+62kfonLxpTen/dW6LsXqFgmr5jhczR5RlB/6v6+ls47\nkPz7oFkiYjB5ZvDO5JKXq1Q4niRJWgGZ1JEkSWqBIhnxz6KVyqjtT56p8qWIuCGldBPQpxjS4Jva\nTah246cv+ebUV5sY36tsf8jJnWqLkCfyt6ib641ihlF9pZh7NzG+1F/t5mBpe58q/UtLi+NIKVWa\nXdVcQ1JKLzexT+ncLblpV7oZusTsk5TSooi4HPg2cAhwTjOPdyfw1YgYRU7irAXcmVJ6NyIepa4c\n2xhaVzKukmprJi0kJyvbUrXSa5Us7WteSelGcpdG9in1VVub5c6U0icAImIN8tpNvwL+FhEfTSm9\n0Yr4mut/N+wjoj95htPpwE3FGlbzq4xr7utxLLBuvW0vAvWTCLW8vktIKf0rIv4K7BMRn04p3VBl\n19eBDSvEQ5F46QL/K4U5r/4+NTqenAT9GU0n1qFlz7/WMcvC+62k9BrMbMa+pb8NDX7PFzNlGszS\njIjPksvyzSXP5JkIvEf+LI4hf/Fj5eYGWySKHyUnKO8lJ4jfBhaRE4IH1XI8SZK0fGnsfwgkSZLU\nTCm7llw3P8jfrIW6m9K13Jyvv7B8ubeLvi1SSl0baZeV7Q9wbhP771JDfP2qrDVSKjX0doW++s+h\nfP/61qm3X1vpqDiac2P0X8V+Y5racYkDR3Qjl3QCOCsiFpc38mLhQW0Lqo8rxuxCXeJmXFnfh4sb\ntmPI12p8LTF3dm10zSspvQ/XbGSffsVjtWTY/6SUZqeULgK+Q17Q/fzWhVe7lNKMlNIZwC+AUcCp\nS+GY61f4/XZFq4Nt6HvkG+xnFe+BSu6neZ/jFieYSlJKT5JnNm4REQc3tX9768j3W7H+XSnRVb9s\nYiWlz9raFY7VncqzBE8nryc2KqW0f0rpOymlU4r1chpdw6eK75CTS19KKY1JKR2TUjq5ON4/WQrv\nGUmS1HmZ1JEkSVq65hSPAVAsBD0BWDsiPrQUjv9Qcexqa33U9wj5m8ItLRFUSTfyAuD17VQ8NnVT\n/4nicduIqPTfozuTEwflx1lEft5Lc3bGE8Uxd6zfERG9gS3J355/dimes7muJa87sXVE7NzYjsW3\n/Ev2JZf9eR64uEqbCGxclApqjnHk12MM+bWZmFKaWvSVZht8iTwj4e4qs7jqK5Xda+vZNu2hLa55\nA0XptmlA74jYtMpupc/lv2s49K/JpRo/GxHVZvO1tZPJZbmOjoj1OiiGmhTrcF1Ift9/o8puvye/\n1w+IiI3aIazvk39nnQ6s2g7na4mOeL/9P/KslkdSSi81Y//S354dKvTtQOX7KMOACSmlJdZiKv7G\nbVth/9Jsumq/AzcoHq+v0LcjzSsjJ0mSllMmdSRJkmoQEZ+PiF0qzVSJiAHkMkyJJWvx/5KcPLiw\n+MZw+ZgoxjXX78nfwj+50g2x4nj/uxGVUpoJXA58JCJOqpREiYhhETGkhhgAzixPJkTEmsBJ5Of+\nh8YGppSmkcvTDCGXSiqPZStgLDmhUV7SaFbxOKjGOBvzZ+AD4Kii1E25HwGrA5ellD5YiudslpTS\nu+S1GQL4S0R8otJ+EbE1OdFXUnr//SCl9LVKDfhxcdxmrVlRvIeeJt+Y3I4ly6s9AMwHTmDJGTxN\nmV3svzRfz6WplhumS/2aN+IPxXHOLmYM/E8xW+qUIpbfN/eARSnJU4rj/riV8bVISmkOcDbQg5zg\n6SxOJSfyf0CFJEqR+DmTvBbK7RHxsSrHqXl9qEpSSq+QZz2tBxzFMnjjvz3fbxHRIyJ+QJ5VNQ/4\nVjOH/p4c3w+KBH/peKsAZ1QZMwXYJCLqz+45nbq17v6nSH7PpvrvwMnF447lGyNiD3IpR0mStAJz\nTR1JkqTabEW+MfR6RPyLvMA35AWh9yTflPxrSum60oCU0sURsS25Bv6LEXEjua7/uuSZD5cAp5Wd\no2pZlZTSmxHxGfK3dx+KiDvJN9wTeSHlrcnlmcpvMH6T/G3yU4GDirinF+ffDPgIOZEyuZnX4DXy\nt54nRMRN5MWg9yeXMft1SulfzTjGEeQSYz8tEhaPkW9ufYb8zfYvp5TeK9v/QfJaBcdERD/q1u/5\nZXFDuGYppSkRcQy5DND4iPgL+XXZgXwdnyF/w7tDpJSuiIgeRXy3R8ST5CTKbPJaSVsDHwJmAETE\nUPJsmjeAGxs59NXkG7/7R8RRKaUmS3WREzkjyO/v/yV1UkoLIuJ+alxPJ6X0XkQ8DGwXEX8GXiC/\n7jc2sYB6ewlgpyplBgHeSimd18bXvJIzyLP09gBeiIhbyQnQdYB9yJ/9y1JKV9Vy0JTSXyLi+8DO\nEbFzSql+cm50RFRLtixOKZ1e07Oo7Hzy2kMHR8RP6s14CGBM/URWmTdTSr+q4VxBfi2qzZ6ZWFbC\nsqqU0hsRcRZ1yYkGSZSU0g8jIpFn0TwUEY+RZ1DOJq/XNZRc2jAB91SJtdpzqOQs8pprG1aKp2xs\nS67ncRHxTpUx41JK91XpW0Ibvd8OjYhdyc+tF3mmy/bka/wK+W/KI82M776I+A3579TTEXEteR2v\nfcm/b2dUGHYueb2gJyPiumL/7YCNgL8Bn6ow5k7y+/Cv5JmjC8mzHe8nz2j6EvDXiLiG/DdvBPAJ\n4C/A55vzXCRJ0vLJpI4kSVJtfka+Ab0LMJJ8g6UHeSbJXcDlKaUr6w9KKR0SEX8nf1P/s+SkyGvk\nm3g31d+9sQBSSuMiYgvywti7kWdQLABeJd8kurbe/nOK2TtfA74A7FfEPJ1c6/8Y8syZ5lpAfv4/\nBj5HXsdjIvDjlNKvK4Vc/zmllCZFxEfIs3v2ICdS3gFuLY7zeL3934qI/cjf4j8Y6Fl0XUZdybvG\nVLymKaXfRMSL5Gu5HzkZ9jLwE+DMlFKlG5it+fZ7TWNTSpcW75tvAruSX7+e5NlaE8gJxtKsjEOL\nxz+llBY2csz3IuJK8o3fg4HzmhHKneSZQ4vJ7/P6fTsDr6eUqpWqq/S8DyTfCN2NfIMyyNd+QtmY\nxq5Xra9DLfsn8g3hamUOJ5Ov29K85k09X1JK8yJiDHA4+ZqNJb8fSmsZXZJS+ksjz6mx458MXEee\npVZeXjGRSxFuWWXcQvJshOaoGkNKaW6RIPl5cbyx9cbtSIVSiYWXyDfUmyuRb9DvW6X/TvLvlvpj\nKjmXfPN//Wr7pJROjogriv12BL5Ift3eAf5Lns15eUqpUtm8auetdq45EXEqjV+Pll7PYxo55gdA\neVKnvd9vhxSPi4B3yUmQO4BbgOtSSu83EksDKaUjI+I58mt2ODlxex35b9azNPybdkFEzCX/Tj6E\n/CWEe8i/575A5aTON8nXbQz5SyFdyLO+7k8pPRkROxXPdU9ymbZ/k5O388h/e5e5mViSJKl9RPNK\nXkuSJEkQEZPIlWOGdXQskiRJkiStaFxTR5IkSZIkSZIkqRMwqSNJkiRJkiRJktQJmNSRJElSrazf\nK0mSJElSB3BNHUmSJEmSJEmSpE6gW0cHsCKIiL7AbsBkYF7HRiNJkiRJkiRJkjpYD2AI8PeU0qzm\nDjKp0z52Ay7v6CAkSZIkSZIkSdIy5YvAFc3d2aRO+5icH/4MbNbkzmuvDdttB9tuCx/9KPTo0aax\nScutY489lnPPPbejw5BWGH7mpPbn505qX37mpPbn505qX37mpPbz7LPPcuCBB8L/8gfNY1KnfcwD\nOOywzXj88VGMH9/4ztOnw7XX5rbKKjBmDHzqU7DnnrDeeu0RrrR86N27N6NGjeroMKQVhp85qf35\nuZPal585qf35uZPal585qUPUtGRLl7aKQg0dcQQ8/ji88gpcdBHssw+sumrjY95/H26+OY9df33Y\ncks46SR46CFYtKh94pYkSZIkSZIkSR3PpE4HGDgQDjsM/vpXmDULbrsNjjwSBg9ueuy//w1nnAFb\nbw0DBsDBB8M118Dbb7d93JIkSZIkSZIkqeOY1OlgPXrA7rvD+efDpEkwYQKcdVZeT6dLE6/OG2/A\nn/4EBxwA/frBzjvDOefACy+0T+ySJEmSJEmSJKn9mNRZhkTA8OHwve/BfffBjBlw+eUwdiyssUbj\nYxcuhLvuguOOg002gY03hmOPhTvvhAUL2id+aVkzduzYjg5BWqH4mZPan587qX35mZPan587qX35\nmZOWfZFS6ugYlnsRMQp4/PHHH2/xQmMLF8KDD+b1dW6+GZ55pvljV1sNdtsN9twT9tgD+vdvUQiS\nJEmSJEmSJGkpGD9+PKNHjwYYnVIa39xxztTpJLp1g+22g5/8BJ5+GiZOhF/9KidrVlqp8bFz5sC1\n18KXv5zX4fn4x+H00+GJJ8CcniRJkiRJkiRJnYNJnU5q6FD45jfh9tth1iy44Qb46ldz0qYxKcHD\nD8MPfwijRsH668Phh8NNN8F777VP7JIkSZIkSZIkqXYmdZYDvXrBvvvC734H06bBY4/BqafCRz/a\n9Nhp0+Cii2CffaBv31ye7YILYMqUto9bkiRJkiRJkiQ1n0md5UyXLjB6dJ6J88gj8NprcMklsN9+\nOfnTmPnz4bbb4MgjYcgQGDkSTjgB7r8fFi1ql/AlSZIkSZIkSVIVJnWWcwMGwFe+AtddB2+8AXfc\nAUcfDcOGNT12wgQ46yzYdlvo3x8OPBCuugpmz277uCVJkiRJkiRJ0pJM6qxAVl4Zdt0VzjsP/vtf\nePZZOPts2GEH6Nq18bFvvgmXXw5jx8Jaa+UxZ5+dj5FS+8QvSZIkSZIkSdKKzKTOCioCNt0Ujj8e\n7r4bZs7Ms3AOPBDWXLPxsYsWwb33wne/C5tvDhtumGf/3HFHLuEmSZIkSZIkSZKWPpM6AmCNNeBz\nn4PLLoMZM/I6OieckNfVacrEifCrX8Fuu0Hfvnn9nksuyev5SJIkSZIkSZKkpcOkjhro2hX+7//g\nxz+G//wHpkyBCy6APfbIJdwa8957cMMN8NWvwrrrwkc/CqeeCo89BosXt0/8kiRJkiRJkiQtj0zq\nqEmDBsHXvw633AKzZsFNN8Hhh8PAgU2PfewxOOWUnNwZODAne264Ad59t83DliRJkiRJkiRpuWJS\nRzXp2RP22gt++1t4+WV44gk4/XT4+MfzOj2Nef31XJZtv/1ymbbddstl2yZObJ/YJUmSJEmSJEnq\nzEzqqMUiYMst4aST4MEHc9LmD3+Az34WVl+98bELFsAdd8DRR8MGG8Dmm8N3vwv33gsLF7ZL+JIk\nSZIkSZIkdSomdbTU9O8PBx8Mf/kLzJwJd94Jxx4LG2/c9Nhnn4Wzz4YddoC11oKxY+Hyy3O5N0mS\nJEmSJEmSZFJHbWSllWDnneGcc+D553M755y8rVu3xse+9RZcdRUceGBOFG27LZx1FkyYACm1T/yS\nJEmSJEmSJC1rTOqoXWy8cZ61c+ed8MYbcM01eVbPWms1Pm7xYrj/fjjhBBg5EoYOhSOPhNtug3nz\n2id2SZIkSZIkSZKWBSZ11O5694bPfCavv/P66/DQQ3ldni23bHrslClwwQWwxx7Qty/ssw9cdBFM\nm9bmYUuSJEmSJEmS1KFM6qhDdekCW20Fp58OTzwBL78MF14Ie+0Fq6zS+Ni5c+Gmm+Dww2G99WDU\nKPjhD+Hhh/MMH0mSJEmSJEmSlicmdbRMWW89+NrXcrJm1iy45Rb4xjdg0KCmxz7xRE4OffzjsM46\n8OUvw3XXwTvvtH3ckiRJkiRJkiS1NZM6Wmatskous/brX8PkyfCf/8CPfwzbbJNn+DRmxoxc3u0z\nn4F+/WCXXeAXv4D//rc9IpckSZIkSZIkaekzqaNOIQJGjoQTToB//QumT4fLLoPPfz6v0dOYDz6A\nO++EY4+FjTaCTTaB446Du+7KfZIkSZIkSZIkdQYmddQp9esHBx4IV14JM2fC3XfD8cfDZps1PfaF\nF+Ccc2DnnfNxDjgA/vSnfBxJkiRJkiRJkpZVJnXU6XXvDjvsAGefDc88k0usnXce7Lpr7mvMO+/A\nNdfAwQfD2mvD1lvDGWfAv/8NKbVP/JIkSZIkSZIkNYdJHS13NtgAjj4a7rgDZs2C66+Hr3wlJ20a\nkxI89BCcdBJsuSUMGgRHHAE33wxz57ZP7JIkSZIkSZIkVWNSR8u11VaDT38aLrkEXn0VHn0UTj4Z\nRo9ueuwrr8CFF8Jee0HfvvCpT8Fvfwsvv9z2cUuSJEmSJEmSVJ9JHa0wunSBj3wETjkFHnssJ3ku\nvhj23Rd69mx87Lx5cMst8PWv5xk8H/oQfP/78MADsGhRu4QvSZIkSZIkSVrBmdTRCmuddeDQQ+GG\nG+CNN+A43IpSAAAgAElEQVT22+Goo2DIkKbH/uc/8OMfwzbb5LJuX/oS/OUv8NZbbR62JEmSJEmS\nJGkFZVJHAnr0gN12g1/+EiZOhKefhp/8BLbfHrp2bXzsrFlw2WXwuc9Bv36w007w85/D88/ndXok\nSZIkSZIkSVoaTOpI9UTA5pvDd78L99wDM2bAFVfAF78Ia6zR+NhFi+Duu+H442HTTWGjjeCYY+Cf\n/4QFC9olfEmSJEmSJEnScsqkjtSENdeEsWPhz3/OCZ777oPvfQ+GD2967EsvwXnnwa675lk8++8P\nv/89TJ/e9nFLkiRJkiRJkpYvJnWkGnTrBttuC2edBRMmwKRJcP75sPvusPLKjY+dMweuvx6+8hUY\nMAA+9jE47TQYP94ybZIkSZIkSZKkppnUkVphyBA48ki47ba8ts6NN8Jhh8G66zY99tFH4eSTYfRo\nWG+9PO7GG+G999o8bEmSJEmSJElSJ2RSR1pKevaEvfeGiy6CV17JM3BOOy3PyGnKq6/CxRfDvvtC\n377wyU/Cr38Nkye3ediSJEmSJEmSpE7CpI7UBiLgwx+GH/wAHn4YXn89r6Wz//7Qq1fjY+fPh9tv\nh29+E4YOhREj4P/9v7yWz8KF7RO/JEmSJEmSJGnZY1JHagdrrw2HHALXXpvLtP3jH3DMMbDBBk2P\nffpp+MlPYPvtoX9/+OIX4cor4c032zxsSZIkSZIkSdIyxKSO1M5WWgl22QXOPRdefBGeew5+9jPY\naSfo1q3xsbNnwxVXwBe+AGutlRM9P/0pPPMMpNQ+8UuSJEmSJEmSOoZJHakDRcAmm8Bxx8G4cTBz\nJlx9NRx0EPTr1/jYxYtzSbbvfQ+GD4dhw+Coo+Dvf4d589onfkmSJEmSJElS+zGpIy1D+vSBAw6A\nP/0pr8PzwANw4omwxRZNj508Gc4/H3bfPSeE9t0XLr4YXn21zcOWJEmSJEmSJLUDkzrSMqprV9h6\nazjjDPj3v2HKFPjNb2DPPaFHj8bHvvce3HgjHHYYDBwIo0fDySfDo4/mGT6SJEmSJEmSpM7HpI7U\nSQwaBEccATffDLNm5ccjjoD11mt67PjxcNpp8LGPwbrrwqGHwvXXw5w5bR+3JEmSJEmSJGnpMKkj\ndUKrrppn7PzmNzB1ap7Jc8YZeWZPRONjp0+HSy+F/feHvn3hE5+AX/4SXnqpfWKXJEmSJEmSJLWM\nSR2pk4vIa+6ceGJeg2f69LwmzwEHwOqrNz72gw/gH/+Ab30LNtwQNtsMvvMduOee3CdJkiRJkiRJ\nWnaY1JGWM2utBQcdBFdfDW+8AePGwXHHwSabND32uefgZz+DHXfMx/n85+HPf87HkSRJkiRJkiR1\nLJM60nKse3fYaaecqHnuOXjhBTj3XNhll9zXmLffzomhgw6CtdeGbbaBM8+Ep56ClNonfkmSJEmS\nJElSHZM60gpko43gmGNyybU33oBrr4VDDoH+/Rsft3hxLu124om51NvgwfCNb8Ctt8L777dL6JIk\nSZIkSZK0wjOpI62gVl8d9t8ffv97eO01ePhh+MEP4MMfbnrsyy/Db34De+4JffvC3nvDhRfCK6+0\nfdySJEmSJEmStKIyqSOJLl3gYx+D006D8eNzcuaii3KyZtVVGx/7/vvwt7/BEUfA+uvnpNBJJ8FD\nD8GiRe0TvyRJkiRJkiStCEzqSGpg4EA47DC48UaYNQtuuw2OPDKXXWvKk0/CGWfA1lvDOuvk8m7X\nXJPX6JEkSZIkSZIktZxJHUmN6tEDdt8dzj8fJk2CCRPgrLNg223zDJ/GzJwJf/wjHHAA9OsHY8bA\nuefCCy+0T+ySJEmSJEmStDwxqSOp2SJg+HD43vfgvvtgxgy4/HIYOxb69Gl87MKFMG4cfPvbsMkm\nsPHG+edx42DBgvaJX5IkSZIkSZI6M5M6klqsb1/4whfgiivyrJx77oHvfAc237zpsS++mGftjBmT\nZ/F89rN5Vs+MGW0ftyRJkiRJkiR1RiZ1JC0V3brB9tvDT38KTz8NL70Ev/wl7LYbrLRS42PnzIFr\nr83r7wwYAB//OPzoR3l9npTaJXxJkiRJkiRJWuYtc0mdiFg5Im6IiOci4omI+HtEDCv67o6IiREx\nvmjfKhu3SkRcEREvFmP3L+uLiPhVRPw3Il6IiCPrnfOkou/FiPhRvb5DizEvRsSFEdG1rO9TEfFs\nRDwfEddGRK+2uzJS5zJsGBx1FNx+O8yaBTfcAIcempM2jUkJHn4YfvAD+PCHYf314fDD4W9/g7lz\n2yd2SZIkSZIkSVoWLXNJncKFKaVNU0ofBm4CLi62J+BbKaVRRTuvbMzxwLyU0kbA7sAFEbFG0XcQ\nsGlKaUNgK+A7EbEZQERsD3wOGAEMB3aLiE8WfUOB04BtiuMOAL5W9PUs4to7pbQJ8Brww7a4GFJn\n16sX7LsvXHwxTJsGjz0Gp5wCH/lI02OnTYOLLoK9987l3vbcEy64AKZMafOwJUmSJEmSJGmZsswl\ndVJK81NKt5dteggYUvbvajF/DvhtcYzJwN3Ap4u+A4DfFX2zgauBsWV9l6WU5qWUFgCXlvXtD9yY\nUppZ/Pu3ZX2fBManlF4s/n1BWZ+kKrp0gdGj4eST4dFH4dVX4ZJL4NOfhp49Gx87bx7ceisceSQM\nGQJbbAEnnAD33w+LFrVL+JIkSZIkSZLUYZa5pE4F3wL+Wvbvn0TEvyPiymImTckgoPy7+5OLbe3Z\nNyAiOsM1lZYZ66wDX/kKXH99LtN2xx1w9NG5fFtTnnoKzjoLtt0W1l4bDjoIrroKZs9u+7glSZIk\nSZIkqb116+gAGhMRJwIbUJQ8Aw5MKU0r+o4EbiaXTOsUjj32WHr37r3EtrFjxzJ2rBN8JICVV4Zd\nd83tF7+A55+Hm2/O7V//anw2zqxZ8Oc/59a1a0707LADDB1a1wYOzH2SJEmSJEmS1F6uvPJKrrzy\nyiW2vf322y061jKb1ImI44F9gTEppXkApYRO8fOvI+JnEbFGUVJtCjAYmF7sMgT4e/Hz1KLv4bK+\nqfX6qNI3rJG+Xcv6hgKvpZQWV3tO5557LqNGjarWLalMBGy6aW7HH59n3/z973DLLbkE25tvVh+7\naBHcc09u5bp1g0GDcum2UqKn9POQITBgQC4PJ0mSJEmSJElLS6XJHePHj2f06NE1H2uZTOpExLeB\nz5MTOnOKbV2BvimlGcW/9wdeLxI6ANcCRwCPFGXZdgC+XvRdAxwWEdcCfcjr7+xZ1nd+RPwKWAx8\nBTi56LsOuC8iTgFmFse/qui7vRi3cUrpheJcpT5JS9kaa8DnP5/bokXw0EN5Bs8tt+QybM2xcCFM\nnJhbJSuvnJM75Yme8uRPv3452SRJkiRJkiRJHWGZS+pExEDgZ8BLwF0REcA8YAxwS0SsBCRykmXv\nsqFnA5dGxH+BhcCRKaXSd/kvAz4CvEhO3PwspfQ0QErpnoi4GphQHPeqlNKtRd+kiDgZeKDouwu4\nsOh7NyK+CtxYJJwmAAe3xTWRtKSuXWGbbXI780yYMiUnd265Be68E+bPb9lx58/PJd+ef75yf8+e\n1RM+Q4dCnz4tfEKSJEmSJEmS1AyRUuroGJZ7ETEKePzxxx+3/JrUxt57D8aNg3/+E156CSZNgsmT\nYe7ctj93796Vy7qVHnv1avsYJEmSJEmSJC37ysqvjU4pjW/uuGVupo4ktUbPnrDXXrmVpAQzZ9Yl\neOo/Tp4MCxa0/txvvw1PPplbJf36VU74DB0KgwdDjx6tj0GSJEmSJEnS8sukjqTlXgT075/bVls1\n7F+8GF57rWHCp/Tz1Kl5HZ/WeuON3B59tHL/OutULus2ZAgMGgTdu7c+BkmSJEmSJEmdl0kdSSu8\nLl1g4MDcttmmYf/ChTBtWuWEz6RJuW9pVLJ87bXcHnywcozrrVe5rNvQoTn2rl1bH4MkSZIkSZKk\nZZdJHUlqQrduuTza4MGV+xcsyLN5qpV3e/311seweHE+x9SpcO+9lWMcNKhhWbfSz2uvnRNDkiRJ\nkiRJkjovkzqS1EorrQQbbphbJXPnwpQp1cu7zZrV+hgWLoSJE3OrpEePnJSqNMtnyJC83k9E6+OQ\nJEmSJEmS1HZM6khSG1t1Vdhss9wqeeednPSpX9at9PjOO62PYd48eP753Crp2bNywqf0c58+rY9B\nkiRJkiRJUuuY1JGkDrb66jByZG71pQRvvVU94TNpErz/futjeO89mDAht0r69Klc1m3IkNx69Wp9\nDJIkSZIkSZIaZ1JHkpZhEbDGGrmNGtWwPyWYObNyWbdJk/IMoAULWh/HW2/Bk0/mVkm/fpUTPkOH\n5rJvPXq0PgZJkiRJkiRpRWdSR5I6sQjo3z+3rbZq2L94Mbz2WuWEz+TJMHUqLFrU+jjeeCO3Rx+t\n3L/OOpXLug0dCuuvD927tz4GSZIkSZIkaXlnUkeSlmNdusDAgblts03D/oULYdq06uXdpk3Ls4Fa\n67XXcnvggcoxrrde9fJuAwdC166tj0GSJEmSJEnq7EzqSNIKrFu3XB5t8GDYcceG/fPnw8svVy/v\nNn1662NYvDjPGJo6Fe69t3KMgwZVL+82YECesSRJkiRJkiQt70zqSJKqWnll2HDD3CqZOzev21Np\nls/kyTBrVutjWLgQJk7MrZIePXJSqlp5t759TfpIkiRJkiRp+WBSR5LUYquuCpttllsl77yTkzv1\nEz6lNmdO62OYNw+efz63Snr1ygmeauXd+vRpfQySJEmSJElSezCpI0lqM6uvDltskVt9KcHs2ZXL\nupUe33+/9TG8+y5MmJBbJX36VC7rVvq5Z8/WxyBJkiRJkiQtDSZ1JEkdIgLWXDO3UaMa9qcEM2dW\nT/hMmQILFrQ+jrfegieeyK2StdaqPstn8OBc/k2SJEmSJElqDyZ1JEnLpAjo3z+3rbZq2L94Mbz2\nWuWEz6RJ8PLLsGhR6+OYOTO3Rx+t3L/OOpUTPkOHwvrrQ/furY9BkiRJkiRJApM6kqROqksXGDgw\nt223bdi/cCG88kr18m7TpuXZQK312mu5PfBA5RjXW69yWbehQ2HddaFr19bHIEmSJEmSpBWDSR1J\n0nKpW7ecPBkyBHbcsWH//Pl5Nk+18m7Tp7c+hsWLYerU3O65p2F/9+4waFD18m4DBuQZS5IkSZIk\nSRKY1JEkraBWXhk23DC3SubOzev21C/rVvr5zTdbH8MHH8BLL+VWSY8edYmpSrN9+vY16SNJkiRJ\nkrQiMakjSVIFq64Km22WWyXvvJMTPJUSPpMmwZw5rY9h3jx47rncKunVq3rCZ+hQ6N279TFIkiRJ\nkiRp2WFSR5KkFlh9ddhii9zqSwlmz65c1q30+P77rY/h3XdhwoTcKunTp3JZt9Jjz56tj0GSJEmS\nJEntx6SOJElLWQSsuWZuo0c37E8JZsxomOgp/TxlCixY0Po43noLnngit0rWWqtywmfo0LzWT48e\nrY9BkiRJkiRJS49JHUmS2lkErL12bltt1bB/8WJ47bXq6/m8/DIsWtT6OGbOzO2RRyr3r7tu5bJu\nQ4bA+utD9+6tj0GSJEmSJEnNZ1JHkqRlTJcuMHBgbttu27B/4UJ45ZXKZd0mTYJXX82zgVrr1Vdz\ne+CByjGuv/6SiZ5hw2D48LwO0SqrtP78kiRJkiRJWpJJHUmSOplu3XISZciQyv3z58PUqdXLu02f\n3voYFi/OZeKmTIF77lmyr0sX2HBDGDkSRoyoe9xwQ+jatfXnliRJkiRJWlGZ1JEkaTmz8sqw0Ua5\nVTJ3bk7G1C/rVnp8883WnX/xYnjhhdyuu65ue48eeRZPebJn5Mhc5i2ideeUJEmSJElaEZjUkSRp\nBbPqqjm5stlmlfvfeadhoqd8ts+cOS0777x58MQTuZVbY42c5Cmf1TNiRN4uSZIkSZKkOiZ1JEnS\nElZfHbbYIrf6UoLZsysnfJ57DiZOrP18s2fDffflVm7gwLrZPKWEz2ab5Rk/kiRJkiRJKyKTOpIk\nqdkiYM01cxs9umH/u+/CM8/AhAnw1FN1jy1Zx2fatNxuv71uW5cuuaxc+ayekSNhgw1cr0eSJEmS\nJC3/TOpIkqSlplcv+NjHcis3c2bDRM+ECTkJVIvFi+H553Orv17P5psvmegZMcL1eiRJkiRJ0vLF\npI4kSWpza60FO+2UW0lKMGVKw2TPc8/BBx/Udvx582D8+NzKrbFGw0TPiBHQp0/rn5MkSZIkSVJ7\nM6kjSZI6RAQMGZLbpz5Vt/2DD+CFFxrO6mnpej333ptbufXWWzLZM3IkbLqp6/VIkiRJkqRlm0kd\nSZK0TOneHYYPz61cab2ep55aMuEzY0bt53jlldxuu61uW9euldfrGTbM9XokSZIkSdKywaSOJEnq\nFKqt1zNjRk7wlJI8pYTPe+/VdvxFi3Lpt+eeg2uvrdu+yiqV1+tZZx3X65EkSZIkSe3LpI4kSerU\n+veHnXfOrWTx4urr9SxcWNvx338fHn88t3JrrtlwVs+IEdC7d+ufkyRJkiRJUiUmdSRJ0nKnSxcY\nOjS3vfaq275gQeX1eiZNqv0cb75Zeb2e9ddvmOjZbDNYeeXWPSdJkiRJkiSTOpIkaYWx0ko5yTJi\nxJLb58ypW6+nvIzbzJm1n+Pll3O79da6baX1esqTPSNH5qST6/VIkiRJkqTmMqkjSZJWeKutBltt\nlVu5GTMazupp7Xo911xTt32VVWD48IZl3AYMcL0eSZIkSZLUkEkdSZKkKvr3hzFjcisprddTf1bP\n88+3bL2exx7LrVzfvg0TPcOHu16PJEmSJEkrOpM6kiRJNShfr2fvveu2L1iQEzvls3qeegomT679\nHLNmwT335FZu0KAly7eNGAGbbup6PZIkSZIkrShM6kiSJC0FK61Ul2wZO7Zu+5w58PTTDWf2vPFG\n7eeYOjW3+uv1bLzxkrN6RoyAYcNyAkqSJEmSJC0/TOpIkiS1odVWg49/PLdy06cvmeQprdczd25t\nx1+0CJ59Nre//KVu+6qrVl6vZ+21Xa9HkiRJkqTOyqSOJElSB1h77dzqr9czefKSs3omTGjZej1z\n58Kjj+ZWrm/fhrN6RoyA1Vdv9VOSJEmSJEltzKSOJEnSMqJLl1w2bdgw2Gefuu3z51der2fKlNrP\nMWsW3H13buUGD244q2eTTVyvR5IkSZKkZYlJHUmSpGXcyivDFlvkVu6dd/J6PeVl3J56KiduajVl\nSm633FK3rVu3vF5PKclTSvgMHep6PZIkSZIkdQSTOpIkSZ3U6qvD1lvnVpLSkuv1lB6ffrr29XoW\nLoRnnsmt0no99cu4uV6PJEmSJElty6SOJEnSciQCBgzIbZdd6rYvXgyTJi05q6e0Xs+iRbWdo9p6\nPf36VV6vZ7XVWv+8JEmSJEmSSR1JkqQVQpcusMEGuVVar6d8Vs+ECS1br+eNN+Cuu3IrN3jwkuXb\nSuv1rLRS656TJEmSJEkrGpM6kiRJK7DG1uuZMKFhGbfWrNdz881127p1y4md8lk9I0fCkCGu1yNJ\nkiRJUjWtTupExDpAf+C/KaX3Wh+SJEmSOtrqq8P//V9uJaX1esqTPKX1et5/v7bjL1yYxz39NFx9\ndd32nj2rr9cjSZIkSdKKrsVJnYjYB/gJsFGxaVdgXET0A/4BnJpS+mvrQ5QkSdKyoHy9nl13rdu+\neDFMnNhwVs8LL9S+Xs9778Ejj+RWbq216pI8pUTP8OGu1yNJkiRJWrG0KKkTEXsB1wMPAlcAp5T6\nUkpvRMQ04MuASR1JkqTlXJcusOGGue27b932efMqr9czdWrt55g5s/J6PUOGNJzV43o9kiRJkqTl\nVUtn6vwQuDeltFNE9KUsqVN4EDi8NYFJkiSpc+vRAz70odzKvf12LrtWKt9WSvi8+Wbt55g8Obe/\n/a1uW2m9nvJZPSNHwuDBrtcjSZIkSercWprUGQF8u5H+6eR1diRJkqQl9O5deb2e119vOKuntev1\nXHVV3fZevXLJtvJZPSNHQn//q1WSJEmS1Em0NKkzF+jZSP8wYFYLjy1JkqQVTASss05un/hE3fZF\ni2DSpIazel58sfb1et59Fx5+OLdy/fs3TPQMH56TQJIkSZIkLUtamtS5Czg4In5RvyMiBgCHATe3\nJjBJkiSpa9e69Xo+/em67fPmwXPPLTmr56mn4OWXaz/HjBkwblxu5YYOrUvylK/X0717656TJEmS\nJEkt1dKkzveBh4BHgWuABOwWETuT19IJ4NSlEqEkSZJUT48esOWWuZV766269XpKiZ6nnoLZs2s/\nx6RJuZWv19O9e916PeUJn0GDXK9HkiRJktT2WpTUSSk9HxHbAucBp5OTON8puu8GjkwpTV4aAUqS\nJEnN1acPbLNNbiUpwWuvLZnkKa3XM29ebcf/4IM8dsKEJbf36pWTPPXLuK21VuufkyRJkiRJJS2d\nqUNK6Wlgl4hYA9gQ6AJMTCnNXFrBSZIkSa0VAeuum1v99XomTlxyVs+ECfDCC7B4cW3nePddeOih\n3Mr1779k+baRI2HzzV2vR5IkSZLUMi1O6pSklGaTy7BJkiRJnUbXrrDRRrntt1/d9nnz4NlnG67X\n88ortZ9jxgy4887cyg0b1nBWz8Ybu16PJEmSJKlxLUrqRMTRwJ4ppd2q9N8G3JRS+k0Ljr0ycBWw\nGfA+MAP4RkrppYhYC/gTsAEwj1zm7b5i3CrAJcBHgUXA91NK1xV9AfwS+CSwGDgvpfTrsnOeBBxC\nXhvo6pTSSWV9hwLfI5eYG1fEsqjo+xRwNnmW0lPAISmld2t9zpIkSVp29OgBH/5wbuXeeqthouep\np/L2Wk2cmNtNN9Vt694dNt10yUTPiBEweHCebSRJkiRJUktn6hxKTnBU8wzwNaDmpE7hwpTS7QAR\ncSRwMbAT8BPgwZTSJyPiI8ANETGkSLIcD8xLKW0UEUOAhyNiXDGT6CBg05TShkW5uCeKvmcjYnvg\nc8AIcsLn/oi4P6V0W0QMBU4DtkwpzYyIG0vPKyJ6FnFtl1J6MSJ+BfwQ+G4Ln7MkSZKWYX36wLbb\n5laSErz6asNkzzPPtGy9nlKiqNxqq8Hw4Usme0aOhH79Wv+cJEmSJEmdS0uTOhsAv26k/zngsJYc\nOKU0H7i9bNNDwHHFz58tzk1K6bGImAbsQE4wfQ74StE3OSLuBj4NXAocAPyu6JsdEVcDY8lJmAOA\ny1JK8wAi4tKi7zZgf+DGsnWCfgucQE5WfRIYn1J6sei7ALgDkzqSJEkrjAgYODC33crmsC9aBC+9\n1HC9nhdfrH29njlzKq/Xs/basPfe8KMf5bV7JEmSJEnLv5YmdRYAAxrpX4c862Vp+Bbw14hYE+iW\nUppR1jcFGFT8PKj4d8nkJvq2Kuu7r17f51p4zAER0SWltLSeuyRJkjqhrl3zGjkbbwz771+3/f33\n4bnnGpZwmzat9nNMnw6/+x3ceitccw1svfXSi1+SJEmStGxqaVLnIeCQiDg3pTSnvCMiegNfLvZp\nlYg4kTwz52vAqq09Xkc79thj6d279xLbxo4dy9ixYzsoIkmSJLWnVVapvF7P7Nk5yVO/jFtz1uuZ\nNg223x7OOQe++U3X35EkSZKkZc2VV17JlVdeucS2t99+u0XHamlS51TgHuDJiPgF8HSxfQRwDHmm\nzhdaeGwAIuJ4YF9gTFEabV5ELIyI/mWzdYYAU4ufpwCDgellfX8vfp5a9D1cYVypjyp9wxrp27Ws\nbyjwWmOzdM4991xGjRpVrVuSJEkrqDXWgO22y62ktF5PaTZP+Xo98+cvOX7hQjj6aHjggTx7p1ev\n9o1fkiRJklRdpckd48ePZ/To0TUfq0tLAkgpPQzsBQRwHvCPov0CSMDeKaUHW3JsgIj4NvB54P+z\nd+fxdo/n/v9fV0ZJRMyUIqnpqKHNzmQm5qmKII1SLa1ycEqrtNoaiiOGNkWp/ChqipDU0GNWESIy\n7phSJVrE6UERYyKJJPf3j8/av73WXluys/baa+3h9Xw87kfW+lzr89nX8k+yvR/Xfe/dYBLoLuCk\n3GcGARuQhUsA44ATc7V+ZGft3JN33w8iolNuG7fhwNi82jER0SMiupOdy3NHrjYeODgi1o2IyD2/\nrvYQ0D8itsi9PymvJkmSJDVL3Xk9++0HP/0p/OlPUFsL8+fD7NkwdGjxPXfcAYMHZ1u8SZIkSZLa\nn5JCHYCU0qPAZsAgYERuDQI2Syk9UupzI2JD4HKgDzAhImZFRF1A9DNgx4h4BbgB+HZKaWmudhnQ\nMyJeBR4ETk4pzcvVbgH+Dswhm9a5PKU0O/c9JpIFPC+STRw9nFJ6IFd7DTgXmAy8QjYFNDpX+xT4\nPnBvrp8NgQtK/d6SJElSU3TuDF/9KjzyCPz858X1l16CQYPgzjsr35skSZIkqWVFSqnaPbR7EVED\nzJw5c6bbr0mSJKms7rsPvvMdaGw75h/9CC69FLp1q3xfkiRJkqQvlrf92oCUUm1T7yv1TB0AIuKr\nZGfOrEG2FVuBlNLNzXm+JEmSpOU7+GCYOROGDYPnniusXXEFTJ+eTe1suGF1+pMkSZIklU9JoU5E\nbArcCgymkTAnJwGGOpIkSVIL23RTeOYZOPlkuPHGwtrkyVBTA2PGwB57VKc/SZIkSVJ5lHqmzmhg\nW+A0oAbo18j6SjkalCRJkrRiPXrADTfA9ddD9+6FtX//G/beG0aOhGXLqtOfJEmSJKn5Sg11dgIu\nSSldlVJ6NqX0RmOrnI1KkiRJWrHjj8+mc/r2Lby+bBn8/Odw6KHw4YdVaU2SJEmS1EylhjrvAY0c\nxSpJkiSp2mpqsnN2DjywuHbffTBwIDz7bOX7kiRJkiQ1T6mhzrXA0RHRuZzNSJIkSSqPNdfMApwL\nL4RocArmP/4BO+wAN91UldYkSZIkSSXqUuJ9rwCdgeci4gbgTWBpww+llP7cjN4kSZIkNUOnTvCL\nXxXuMacAACAASURBVMDgwXDUUfDee/W1hQvhe9/Ltmq78kpYZZXq9SlJkiRJappSQ52xea8v/4LP\nJLLgR5IkSVIV7b031NbCEUfA1KmFteuuy7ZqGzcO+vWrTn+SJEmSpKYpNdQZWtYuJEmSJLWojTaC\nJ5+En/wEfv/7wlptLQwYALfeCgccUJ3+JEmSJEkrVlKok1KaWO5GJEmSJLWsbt3gqquy83R+8ANY\nsKC+9sEHcOCB8KtfwbnnQmdn7iVJkiSp1elU7QYkSZIkVdZRR8G0abDFFsW1Cy6A/fcvPH9HkiRJ\nktQ6lLr9GhGxPnA8UAP0oTggSimlPZvRmyRJkqQWsvXWMH06HH98dp5OvkcfhZoauOsuGDKkOv1J\nkiRJkoqVNKkTEdsBfwN+CWxKdsbOOsDmwO7ARkCUp0VJkiRJLWG11eDOO+G3vy3ebu3NN2GXXeCa\nayCl6vQnSZIkSSpU6vZrI4FPgS2BvcgCnB+llDYChgNrAD8rS4eSJEmSWkwEnH46TJgAX/pSYe3z\nz+Hkk+GYY2D+/Or0J0mSJEmqV2qosxMwOqU0F1iW/6yU0l3AbcBlzW9PkiRJUiXssgvU1sJuuxXX\nbrst24btlVcq35ckSZIkqV6poU4n4J3c6w+BpcCaefUXgAHN6EuSJElSha2/Pjz2GJx5ZnFt9mwY\nOBDGj698X5IkSZKkTKmhzmtAP4CU0rLc+73y6juShT2SJEmS2pAuXeCSS+DPf87O3Mn3ySdw+OFw\nxhnZ1mySJEmSpMoqNdR5BDgi7/0fgO9HxGMR8VfgWOD25jYnSZIkqToOPRRmzIBtty2u/eY3sOee\n8NZble9LkiRJkjqyUkOdi4AREdE19/53wDnAWkAf4ALgl81vT5IkSVK1bL45TJkCxxxTXHvqKejf\nHyZOrHxfkiRJktRRlRTqpJQ+SCnNTCl9nnufUkoXppT6p5QGppTOSyktLm+rkiRJkiqtZ0/405/g\n2muhW7fC2jvvZBM7l10GKVWnP0mSJEnqSEqd1JEkSZLUQUTAD38ITz8Nm2xSWFu6FM48E4YNg48+\nqk5/kiRJktRRdGnKhyLiBiABJ6SUluber0hKKR3frO4kSZIktRoDB8LMmXD00fDQQ4W1u++GF16A\n8eNhu+2q058kSZIktXdNCnWAPYBlZJM9S3PvV7TBghswSJIkSe3MWmvB/ffDhRfCeecVbrv26quw\n/fYwenTj5/BIkiRJkpqnSaFOSqnv8t5LkiRJ6jg6dYJzzoEhQ+Coo2DevPraZ5/Bd74DkyfD734H\n3btXr09JkiRJam9W+kydiFglIv4rInZtiYYkSZIktQ377gu1tTBoUHHt2mthl13gjTcq35ckSZIk\ntVcrHeqklBYClwBblr8dSZIkSW3JJpvAU0/BSScV16ZPh5oaePjhyvclSZIkSe3RSoc6OS8CfcvY\nhyRJkqQ2qnt3uOYauPlm6NGjsDZvHuy/P5x/PixbVp3+JEmSJKm9KDXU+QXww4jYq5zNSJIkSWq7\njjkGpk6FzTYrvJ4SnHceHHggvP9+VVqTJEmSpHah1FDnFGAe8HBEvBoRj0bEfQ3WvWXsU5IkSVIb\nsO22MGMGHHpoce2hh7Lt2KZPr3xfkiRJktQelBrqbAd0BeYCnYHNgG0bWZIkSZI6mD59YPx4uOwy\n6Ny5sDZ3Luy8M4wenU3wSJIkSZKarkspN6WU+pa5D0mSJEntSASccQYMGgTDh8M779TXFi+GE0+E\nyZPhD3+Anj2r16ckSZIktSWlTupIkiRJ0grtthvMmpVN5zR0882www4wZ07l+5IkSZKktqjZoU5E\n9I6IL0fExg1XORqUJEmS1LZ96Uvw+OPwk58U155/HgYOhHvuqXxfkiRJktTWlBzqRMRJETEH+BB4\nA3itkSVJkiRJdO0Kl18Od90FvXsX1j7+GA49FM46C5YsqU5/kiRJktQWlBTqRMSJwNXAq8AvgQB+\nB4wE3gaeA44vU4+SJEmS2onDD4fp02HrrYtrl14Ke+8Nb79d+b4kSZIkqS0odVLnVODhlNL+wP+X\nu3Z/SukXwFeB3sBaZehPkiRJUjuz5ZYwdSocdVRx7YknoKYGJk2qeFuSJEmS1OqVGupsCvwl9/rz\n3J/dAFJKHwHXA//ZvNYkSZIktVe9esGtt8LVV2dbs+V76y3YfXcYNQpSqkp7kiRJktQqlRrqfAR0\nAUgpfQwsADbKq38CrN+81iRJkiS1ZxHwn/8JTz0FG21UWFu6FH78YzjyyOzMHUmSJElS6aHOi8DX\n8t5PAU6KiA0jYiPgh8ArzW1OkiRJUvs3ZAjU1mbn6TQ0bhwMHgyzZ1e+L0mSJElqbUoNdW4FtomI\n7rn35wJbAXOB14EtgV82uztJkiRJHcLaa8ODD8I55xTXXn45C3Zuv73yfUmSJElSa1JSqJNSujGl\nNCSltCj3/mlga+DHwI+A7VJK95evTUmSJEntXefOcP75cP/9sMYahbUFC+Db34ZTT4XFi6vTnyRJ\nkiRVW6mTOkVSSv9MKV2RUvp9Ssmt1yRJkiSV5IADsu3YBgworv3+97DrrvDmm5XvS5IkSZKqraRQ\nJyKmRcTpEfHlcjckSZIkSX37wqRJcMIJxbWpU6GmBh57rOJtSZIkSVJVlTqpsxT4DfB6REyKiFMi\nYv0y9iVJkiSpg1tlFRg9Gm68MXud7733YJ994KKLYNmy6vQnSZIkSZVW6pk6OwB9gZ8D3YErgTcj\n4vGIOCEi1i5fi5IkSZI6su9+F555BjbdtPB6SvDLX8LBB8MHH1SlNUmSJEmqqJLP1EkpzU0pXZZS\nGgRsBpwDrAFcC/xfRDxUph4lSZIkdXBf/zrMmJEFOA3df3+2HVttbeX7kiRJkqRKKjnUyZdS+mdK\n6WKgBvgh8BmwdzmeLUmSJEkAq68Od98NI0dCpwa/ybz+Ouy4I/zxj1VpTZIkSZIqoiyhTkRsHxG/\nBeaSTeoA3F6OZ0uSJElSnU6d4Kyz4LHHYN11C2uLFsH3vw/HHQeffVad/iRJkiSpJZUc6kTEgIi4\nNCJeB54GTsj9eTiwbkrpmPK0KEmSJEmFhg7Ntlvbccfi2o03Ztf/8Y/K9yVJkiRJLamkUCci/gFM\nA04FaoGjyIKcb6WU7k4pLSpjj5IkSZJUZMMN4Ykn4LTTimvPPgsDBsBf/lLxtiRJkiSpxZQ6qfM3\n4FiyIOewlNLYlNKCMvYlSZIkSSvUtSuMGgVjx0KvXoW1jz6Cgw+Gs8+GJUuq058kSZIklVNJoU5K\n6RsppVtTSp+UuyFJkiRJWllHHgnTp8NWWxXXLr4Y9t0X/v3vyvclSZIkSeVU8pk6kiRJktSabLUV\nTJsGw4cX1x5/HGpqYPLkyvclSZIkSeViqCNJkiSp3Vh1VRgzBq68Erp0Kaz961+w225ZLaXq9CdJ\nkiRJzWGoI0mSJKldiYBTT4WJE2HDDQtrS5bAj34EI0bAp59Wpz9JkiRJKpWhjiRJkqR2accdobYW\n9tijuDZ2LAweDC+9VPm+JEmSJKlUhjqSJEmS2q1114VHHoGzzy6uvfQSDBoEd95Z+b4kSZIkqRSG\nOpIkSZLatc6d4aKL4L77oE+fwtr8+TB8OJx2GixeXJ3+JEmSJKmpuqz4IxARrwEre5RoSiltuvIt\nSZIkSVL5feMb2XZsw4bBs88W1q64AqZPz6Z2Gp7DI0mSJEmtRVMndSY2suYDfYEFwKzcWpC79inw\nZHlblSRJkqTm+cpXYPJkOO644trkydC/Pzz+eOX7kiRJkqSmaFKok1L6bkrpe3ULuBf4MrB3Smmb\nlNKw3NoG2BfYCLinlIYi4oqIeC0ilkXEdnnXn4iIf0ZEbW79KK/WIyJuj4g5EfH3iBiWV4uIuCoi\nXo2IVyLi5AY/75e52pyIuLBB7fjcPXMiYnREdM6rHRQRL0XEyxExLiJWLeX7SpIkSaqsHj3gj3+E\n66+H7t0La+++C3vvDSNHwrJl1elPkiRJkr5IqWfq/Bq4KqX014aFlNKjwO+BC4vuapq7gJ2A1xs+\nGvhRSqkmt67Iq50BLEwpbQ7sB1wTEWvkascA/5FS2gwYAvw0IrYCiIhdgeHANsDWwL4RsX+u1i/3\nPXfKPXd94IRcrRdwPXBwSmlL4C3gnBK/ryRJkqQqOP74bDqnX7/C68uWwc9/DoceCh9+WJ3eJEmS\nJKkxpYY6mwPvL6f+PlDSeToppUkppf8DopHyF/U7HLg2d//rwBPAobnakcB1udoHwFhgRF7tlpTS\nwpTSYuCGvNow4N6U0ru599fm1fYHalNKc3Lvr8mrSZIkSWojampg5kw48MDi2n33wYABxefvSJIk\nSVK1lBrq/AP4XmNbjkVEb+A44J/NaewLXBIRz0XEmNwkTZ2NgTfy3r+eu1bJ2voRUep/T0mSJElV\nssYaWYBz0UXQqcG/6P/5T9hhB7jppqq0JkmSJEkFupR43y+BccDfI+Im4NXc9c2BY4H1gCOa3V2h\no1NK/wLInYvzP2RbprUZp59+On369Cm4NmLECEaMcMhHkiRJqqZOneDss2HwYBgxAt57r762cCF8\n73vZVm1XXgmrrFK9PiVJkiS1PWPGjGHMmDEF1z766KOSnlVSqJNSuiciDgAuAc5uUH4WOD6l9HBJ\nHX3xz/xX3uurI+LyiFgjt6XaG8AmwDu5j/QF6n7+3Fxtal5tboMaX1D7ynJqe+fV+gFvpZSWe5Tq\nqFGjqKmpWd5HJEmSJFXRXntBbS0ceSRMmVJYu+66bKu2ceOKz+GRJEmSpC/S2HBHbW0tAwYMWOln\nlbxdWErpkZRSf2ADYIfc2iClVFPuQCciOkfEunnvhwFv5wIdyKaGTszV+gG7AffkancBP4iIThGx\nJtn5O2PzasdERI+I6E62bdwdudp44OCIWDciIvf8utpDQP+I2CL3/qS8miRJkqQ2bKONYOJEOPXU\n4lptbXbOzgMPVL4vSZIkSWr2GTAppbdTSlNz6+3mPi8iro2IN4ENgYcj4hWgG3B/7jydZ8kCloPz\nbrsM6BkRrwIPAienlOblarcAfwfmkE3rXJ5Smp3rfSJZwPMiMBt4OKX0QK72GnAuMBl4hWwKaHSu\n9inwfeDeXH8bAhc097tLkiRJah26dcu2Wrv9dujZs7D2wQdw4IFwzjmwdGl1+pMkSZLUMUVKqbQb\nIzYm23ptKLAOcEhK6cmIWBs4B7gxpTSrbJ22YRFRA8ycOXOm269JkiRJbczs2TBsGLz8cnFt772z\n4GfttSvflyRJkqS2K2/7tQEppdqm3lfSpE5EfBWYRbaV2WtAH3Ln86SU3gN2Bk4p5dmSJEmS1Jps\nvTVMnw6HH15ce/RRqKmBqVOLa5IkSZJUbqVuv3Yp8CGwBXA0EA3q9wO7NKMvSZIkSWo1eveGO++E\nUaOgS5fC2ptvwi67wDXXQIkbIUiSJElSk5Qa6uwK/CGl9C7Q2K8tc8nOmZEkSZKkdiECTjsNJkyA\nL32psPb553DyyXDMMTB/fnX6kyRJktT+lRrqdAIWLKe+DrCoxGdLkiRJUqu1885QWwu7715cu+02\nGDKk8fN3JEmSJKm5Sg11aoEDGytERBfgW8CUUpuSJEmSpNZs/fWz83TOOqu4Nns2DBoE48dXvi9J\nkiRJ7Vupoc7FwH4R8Qdgm9y19SJiL+ARYCtgZBn6kyRJkqRWqUsXGDkS7r4bVlutsPbJJ3D44XDG\nGdnWbJIkSZJUDiWFOimlB4HvAsOBx3OXbyULdGqA76SUnixHg5IkSZLUmh1yCMycCdttV1z7zW9g\nzz3hrbcq35ckSZKk9qfUSR1SSrcAGwGHA2cBZwNHAhullMaUpz1JkiRJav022wyeeQa+853i2lNP\nQf/+MHFi5fuSJEmS1L6UHOoApJTmp5TuTildllK6JKU0LqX0SbmakyRJkqS2omdPuOkmGD0aunUr\nrL3zTjaxc9llkFJV2pMkSZLUDnRpzs0R0RvYBFgDiIZ1t2CTJEmS1JFEwAknQE1NdqbOG2/U15Yu\nhTPPhMmTs/CnT5+qtSlJkiSpjSppUici1oqIMcD7wHPAE7k1IbfqXkuSJElShzNwYHbOzn77Fdfu\nuSerP/985fuSJEmS1LaVuv3adWRn6VwFDAP2AIbm/sx/LUmSJEkd0lprwf33w/nnZxM8+V59Fbbf\nHm65pTq9SZIkSWqbSt1+bR9gVErpzHI2I0mSJEntSadOcM45MGQIHHUUzJtXX/vsM/jOd7Lt2H73\nO+jevXp9SpIkSWobSp3UWQC8XsY+JEmSJKnd2ndfqK2FQYOKa9deCzvvXHj+jiRJkiQ1ptRQ51bg\n0HI2IkmSJEnt2SabwFNPwUknFddmzICaGnjoocr3JUmSJKntKDXUGQesGREPRcRhETEoImoarnI2\nKkmSJEltXffucM01cPPN0KNHYW3ePDjggOwMnmXLqtOfJEmSpNat1DN1JuW93ruRegAJ6Fzi8yVJ\nkiSp3TrmGPj612HYMJgzp/56SnDeeTBlCtx6K6y1VtValCRJktQKlRrqfK+sXUiSJElSB7PttjB9\nOnzve3D33YW1hx7KtmMbN67xc3gkSZIkdUwlhToppT+VuxFJkiRJ6mj69IHx4+E3v4Gf/QyWLq2v\nzZ0LO+8MV14JJ5wAEdXrU5IkSVLrUOqZOpIkSZKkMoiAM86Av/4V1luvsLZ4MZx4Inz3u7BgQVXa\nkyRJktSKNGlSJyJuIDsj54SU0tLc+xVJKaXjm9WdJEmSJHUQu+0Gs2bB8OHw1FOFtZtvzmrjx8Pm\nm1enP0mSJEnV19Tt1/YAlpFN9izNvU8ruGdFdUmSJElSni99KZvYOftsuPzywtoLL8DAgfCnP8Eh\nh1SnP0mSJEnV1aTt11JKfVNKX0kpfZ73vt8K1ldatnVJkiRJan+6doXLLoNx46B378Laxx/DoYfC\nWWfBkiXV6U+SJElS9XimjiRJkiS1QsOGwYwZsM02xbVLL4W99oK33658X5IkSZKqx1BHkiRJklqp\nLbaAKVPg6KOLaxMnQk0NTJpU+b4kSZIkVUfJoU5E7B8Rj0bE+xGxJCKWNlzlbFSSJEmSOqJeveDm\nm+Gaa7Kt2fK99RbsvjuMGgXJU00lSZKkdq+kUCcihgH/A6wH3JF7zpjc68+A54Ffl6lHSZIkSerQ\nIuCkk7KpnI02KqwtXQo//jEceWR25o4kSZKk9qvUSZ2fA9OA/sC5uWs3pJS+DWwDfAl4rfntSZIk\nSZLqDB4MtbWwzz7FtXHjsvrs2ZXvS5IkSVJllBrqfBW4I6W0FFiSu9YVIKX0OnANcFazu5MkSZIk\nFVh7bXjgATjnnOLayy9nwc5tt1W+L0mSJEktr9RQZwGwGCCl9CGwiGw6p847QL/mtSZJkiRJakzn\nznD++XD//bDGGoW1BQvg6KPhlFNg0aLq9CdJkiSpZZQa6rxMNq1T51ngmIjoEhGrAEcBc5vbnCRJ\nkiTpix1wQLYd24ABxbWrr4bddoM336x8X5IkSZJaRqmhzt3ANyOie+79RcDuwIfAu8AuwMhmdydJ\nkiRJWq6+fWHSJDjhhOLa1KnQvz88+mjF25IkSZLUAkoKdVJKl6eUNk4pLcq9/x+yUOc6YDSwZ0rp\npnI1KUmSJEn6YqusAqNHw403Zq/zvf8+7LsvXHghLFtWnf4kSZIklUepkzpFUkpPpZROTymdkVKa\nUK7nSpIkSZKa5rvfhSlTYNNNC6+nBL/6FXzjGzBvXlVakyRJklQGZQt1JEmSJEnV97WvwYwZcPDB\nxbUHHsjO35k5s/J9SZIkSWq+JoU6EfFaRPxzJdc/Wrp5SZIkSVKx1VeHu++GkSOhU4Pf+l5/HXba\nCa6/PpvgkSRJktR2dGni5yYC/nNfkiRJktqITp3grLNg8GD41rfg3/+ury1aBD/4AUyeDFdfDT16\nVK9PSZIkSU3XpFAnpfTdFu5DkiRJktQChg6F2lo48sgsxMl3441Zbfz44nN4JEmSJLU+nqkjSZIk\nSe3chhvCE0/AaacV1557Ljtn5777Kt6WJEmSpJVUcqgTEd0j4pSIeCAi/pZbD+SurVLOJiVJkiRJ\nzdO1K4waBWPHwqqrFtY++gi++U04+2xYsqQ6/UmSJElasZJCnYj4MvAscCXwNeDd3Ppa7tqzuc9I\nkiRJklqRI4+EadNgq62KaxdfDPvuW3j+jiRJkqTWo9RJnauBTYAjU0obppR2y60NgeHAxrnPSJIk\nSZJama22yoKdESOKa48/Dv37F5+/I0mSJKn6Sg119gRGpZTGNSyklO4Crsh9RpIkSZLUCq26Ktx2\nG1x1VbY1W77/+z/YbTe48kpIqTr9SZIkSSpWaqjzCbC8gfy3c5+RJEmSJLVSEXDKKTBxImy4YWFt\nyRL40Y+yaZ5PP61Of5IkSZIKlRrq3Ah8NyJ6NixExKrA94A/NqcxSZIkSVJl7LADzJoFezay38LY\nsTB4MLz0UuX7kiRJklSo1FDn2dy9f4+ICyLi2Ny6EPhb7jPPR8Rh+assHUuSJEmSym6ddeDhh+EX\nvyiuvfQSDBqUBTySJEmSqqdLiffdkfe6kX/y82VgDBB51xLQucSfJ0mSJElqYZ07w4UXwvbbwzHH\nwIcf1tfmz4dvfQueeQYuvRS6daten5IkSVJHVWqoM7SsXUiSJEmSWo2DDoKZM+Hww7Nt2fJdcQVM\nnw533ll8Do8kSZKkllVSqJNSmljuRiRJkiRJrcdXvgJPPw2nngp/bHBi6uTJ0L8/jBnT+Dk8kiRJ\nklpGSWfqRMS2TfjM4aU8W5IkSZLUOvToAddfn63u3Qtr774L++wDF18My5ZVpz9JkiSpoykp1AFm\nRMTPI6Lo/ohYMyLGAh6hKUmSJEntwPHHZ2fp9OtXeH3ZMjj7bDjkEPjgg+r0JkmSJHUkpYY6fwIu\nAiZHxJZ1FyPiEGA2cCBwWvPbkyRJkiS1Bv37Z+fsHHRQce0vf4GBA+HZZyvflyRJktSRlBTqpJRO\nAPYHvgzMioizIuJW4M/AP4Cvp5SuKl+bkiRJkqRqW2MNuPdeuOgi6NTgt8l//hN22AFuvLE6vUmS\nJEkdQamTOqSUHga2Bp4H/hsYQTa9s0tK6dXytCdJkiRJak06dcq2XHv4YVh77cLawoVw3HHwgx9k\nryVJkiSVV8mhTkT0Ai4FBpMFO58Bx5FN8EiSJEmS2rG99oJZs2D77Ytr118PO+0Er71W+b4kSZKk\n9qykUCcihgIvAMcCPwcGAP2B14G/RMT1EdG7XE1KkiRJklqfL38ZJk6EU08trtXWQk0N3H9/5fuS\nJEmS2qtSJ3UeAz4ABqSULkkpLUspzQF2Bs4CjiILfSRJkiRJ7Vi3bnDllXD77dCzZ2Htww/hoIPg\nV7+CpUur058kSZLUnpQa6lwADEkpzc6/mDKXk03uvNPc5iRJkiRJbcOIETBtGmy5ZXHtwgthv/3g\n3Xcr35ckSZLUnpQU6qSUzkspLVlO/SVgh5K7kiRJkiS1OVtvDdOnwxFHFNceeyzbjm3KlMr3JUmS\nJLUXTQ51ImJwRKzZxM/2A44upaGIuCIiXouIZRGxXd71dSLiwYh4JSKej4hd8mo9IuL2iJgTEX+P\niGF5tYiIqyLi1dy9Jzf4eb/M1eZExIUNasfn7pkTEaMjonNe7aCIeCkiXo6IcRGxainfV5IkSZLa\nk969YexYGDUKunQprP3v/8Kuu8LVV0NK1elPkiRJastWZlLnGWC/ujcRsWZELIiI3Rr57I7AjSX2\ndBewE/B6g+sjgWdSSlsAxwG354UsZwALU0qb53q8JiLWyNWOAf4jpbQZMAT4aURslfsOuwLDgW2A\nrYF9I2L/XK0f8Gtgp9xz1wdOyNV6AdcDB6eUtgTeAs4p8ftKkiRJUrsSAaedBk88ARtsUFj7/HM4\n5RQ4+miYP78q7UmSJElt1sqEOtHI+1WAzo18tmQppUkppf9r5OcdCVyb+8wM4F9AXaA0PK/2OvAE\ncGjefdflah8AY4ERebVbUkoLU0qLgRvyasOAe1NKdbs+X5tX2x+oTSnNyb2/Jq8mSZIkSQJ22glq\na2Ho0OLa7bfDkCHw8suV70uSJElqq0o6U6fSctu+dUkp/Tvv8hvAxrnXG+fe13m9CrX1I6JN/PeU\nJEmSpEpZbz145BH42c+Ka7Nnw8CBMG5c5fuSJEmS2qIuK/6IyuX000+nT58+BddGjBjBiBEO+UiS\nJElqv7p0gYsvhu23h2OPhY8+qq99+ikccQT8+McwciR07Vq9PiVJkqSWMGbMGMaMGVNw7aP8fxSv\nhDYR6qSU5kXEkohYN29apy8wN/f6DWAT4J282sO513NztamN3FdX4wtqX1lObe+8Wj/grZTSsuV9\nj1GjRlFTU7O8j0iSJElSu/XNb8KMGTBsGDz/fGHtt7+FadNg7Njic3gkSZKktqyx4Y7a2loGDBiw\n0s9a2e3C+kZETUTUANvlrm1edy2v1m+lO1mxu4CTACJiELABMDFXGwecmKv1Iztr5568+34QEZ1y\n27gNJztXp652TET0iIjuwHHAHbnaeODgiFg3IiL3/LraQ0D/iNgi9/6kvJokSZIk6Qtsthk880w2\nsdPQpElQUwMTJxbXJEmSJK38pM4FuZXvmkY+F0AqpaGIuBY4EFgPeDgiPkkpbQH8DLglIl4BFgHf\nTiktzd12GXBDRLwKLAFOTinNy9VuAQYCc4BlwOUppdkAKaWJETEWeDHX7x0ppQdytdci4lxgcq42\nARidq30aEd8H7o2Izrn7G/mVRJIkSZLUUM+ecOONsNNOcMopsHhxfe2dd2DPPbPt2s44AyKq16ck\nSZLU2kRKTcteImKlQ4uU0p9WuqN2KDe9NHPmzJluvyZJkiRJeWbMgMMPhzfeKK4dcgjcdBM0OJpU\nkiRJavPytl8bkFKqbep9TZ7UMaCRJEmSJJXbwIFQWwtHHw0PPlhYu+eerD5+PGy3XeP3S5Ik+V8B\nyQAAIABJREFUSR3Jyp6pI0mSJElSWa25JvzP/8Cvf1283dqrr8L228PNN1enN0mSJKk1MdSRJEmS\nJFVdp07wq1/BQw/BWmsV1j77DI49Fk48ERYurE5/kiRJUmtgqCNJkiRJajX22Sfbjm3w4OLa6NGw\nyy7w+usVb0uSJElqFQx1JEmSJEmtysYbw5NPwn/+Z3FtxgwYMCCb6JEkSZI6GkMdSZIkSVKr0707\nXH013HIL9OhRWJs3Dw44AM47D5YurUp7kiRJUlUY6kiSJEmSWq2jj4apU2HzzQuvpwTnnw8HHgjv\nvVed3iRJkqRKM9SRJEmSJLVq226bbbt22GHFtYcfzrZjmz698n1JkiRJlWaoI0mSJElq9VZbDcaN\ng8svh86dC2tz58LOO8O112YTPJIkSVJ7ZagjSZIkSWoTIuAnP4HHH4f11y+sLV4MJ50Exx4LCxZU\npz9JkiSppRnqSJIkSZLalF13hVmzsj8buuUW2H57mDOn8n1JkiRJLc1QR5IkSZLU5qy/Pvz1r/DT\nnxbXXngBBg6Eu++ufF+SJElSSzLUkSRJkiS1SV26wKWXwvjx0Lt3Ye3jj+Gww+DMM2HJkur0J0mS\nJJWboY4kSZIkqU077DCYMQO22aa4dtllsOee8Pbble9LkiRJKjdDHUmSJElSm7fFFjBlChx9dHHt\nySehf3946qnK9yVJkiSVk6GOJEmSJKld6NULbr4Z/vAH6NatsPb22zB0KPz2t5BSdfqTJEmSmstQ\nR5IkSZLUbkTAiSdmUzkbb1xYW7oUfvITOOKI7MwdSZIkqa0x1JEkSZIktTuDB8PMmbDPPsW18eNh\n0CB48cXK9yVJkiQ1h6GOJEmSJKldWntteOABOPfcbIIn3yuvwJAhcNtt1elNkiRJKoWhjiRJkiSp\n3ercGc47D+6/H9Zcs7C2YAEcfTScfDIsWlSV9iRJkqSVYqgjSZIkSWr39t8/245twIDi2jXXwK67\nwty5le9LkiRJWhmGOpIkSZKkDqFvX5g0CX74w+LatGlQUwOPPFLxtiRJkqQmM9SRJEmSJHUYq6wC\n114LN92Uvc73/vuw335wwQWwbFlV2pMkSZKWy1BHkiRJktThHHssTJkCm25aeD0lOOccOOggmDev\nOr1JkiRJX8RQR5IkSZLUIX3tazBjBnzzm8W1Bx/MtmObObPyfUmSJElfxFBHkiRJktRhrb463H03\nXHIJdGrwG/Ibb8COO8J112UTPJIkSVK1GepIkiRJkjq0CDjzTPjrX2HddQtrixfDCSfAccfBggXV\n6U+SJEmqY6gjSZIkSRKw++4waxbstFNx7aabsqmdcePg3Xcr3ZkkSZKUMdSRJEmSJClngw1gwgT4\n8Y+La889B0cckU3zbLstnHoqjB9vyCNJkqTKMdSRJEmSJClP167wm9/AXXfBqqs2/pkXX4Tf/x4O\nP7ww5Pnzn+G99yrbryRJkjoOQx1JkiRJkhpx+OEwYwZ89asr/mxdyDNsGKyzDmy3HfzXfxnySJIk\nqby6VLsBSZIkSZJaqy23hOnT4frr4YEHYNIkmD9/xfe98EK2rroqe7/tttmZPUOHwq67wlprtWjb\nkiRJaqcipVTtHtq9iKgBZs6cOZOamppqtyNJkiRJKtHnn8PMmfDEE9lqasjT0HbbZSHP7rsb8kiS\nJHVEtbW1DBgwAGBASqm2qfcZ6lSAoY4kSZIktU91Ic+ECfUhz4IFK/eMiOKQZ801W6BZSZIktRqG\nOq2YoY4kSZIkdQyff56dw5M/yWPII0mSpIYMdVoxQx1JkiRJ6pgWLy4MeZ5+urSQ52tfKwx51lij\n/L1KkiSpcgx1WjFDHUmSJEkSFIc8kybBZ5+t3DPyQ56hQ2GXXQx5JEmS2hpDnVbMUEeSJEmS1JjF\ni2H69MJJnlJCnq9/vX6Sx5BHkiSp9TPUacUMdSRJkiRJTVEX8kyYkIU8kyeXFvL0718Y8qy+egs0\nK0mSpJIZ6rRihjqSJEmSpFIsWlQ8ybNw4co9w5BHkiSp9THUacUMdSRJkiRJ5bBoEUybVh/yTJ68\n8iFPp07FIU+fPuXvVZIkSV/MUKcVM9SRJEmSJLWEcoc8Q4fCzjsb8kiSJLU0Q51WzFBHkiRJklQJ\nCxcWhzyLFq3cMzp1gpqa+kkeQx5JkqTyM9RpxQx1JEmSJEnVkB/yTJgAzzxTesgzdGh9yLPaai3R\nrSRJUsdhqNOKGepIkiRJklqDhQth6tT6SZ5SQ54BAwoneQx5JEmSVo6hTitmqCNJkiRJao0WLoQp\nUwpDnsWLV+4ZnTsXhzy9e5e/V0mSpPak1FCnS8u1JEmSJEmSWrNVVqkPYwA++6x4kmdFIc/SpdkW\nb9OmwaWXGvJIkiS1JEMdSZIkSZIEQI8exSFP/iTPlCmlhTwDB9Y/d6edDHkkSZJKZagjSZIkSZIa\n1aMHDB2aLSgMeSZMyF5//vnyn7F0aTb9M3UqXHJJYcgzdGgW8qy6akt/E0mSpPbBM3UqwDN1JEmS\nJEnt0YIFxZM8Kwp5GurcGQYNKpzkMeSRJEntnWfqSJIkSZKkiurZE/bYI1tQH/JMmJCFPFOnNm2S\nZ8qUbI0cCV26ZJM8Q4dmIc+OOxrySJIk1THUkSRJkiRJZdFYyPPMM/WTPE0JeZYsqQ95Lr44C3ka\nTvL06tWy30OSJKm1MtSRJEmSJEktomdP2HPPbAHMn18Y8kyb1rSQ55lnslUX8gweXB/y7LijIY8k\nSeo4DHUkSZIkSVJF9OoFe+2VLSgOeaZOzUKc5VmyBCZPztZ//7chjyRJ6lgMdSRJkiRJUlU0FvJM\nnlw4ybOyIU/XrsUhT8+eLfs9JEmSKsVQR5IkSZIktQq9esHee2cL6kOeCROykGf69BWHPJ9/Dk8/\nna2LLioMeYYOhR12MOSRJEltl6GOJEmSJElqlRqGPJ9+WjjJU2rIM2RI/SSPIY8kSWpLDHUkSZIk\nSVKbsOqqsM8+2YIs5Hn66cKQZ+nS5T/j889h0qRsXXghdOtWHPL06NGy30OSJKlUhjqSJEmSJKlN\nWnVV2HffbAF88knxJM+KQp7Fi+Gpp7J1wQWGPJIkqXUz1JEkSZIkSe1C797FIU/+JM+MGaWFPNtv\nXx/ybL+9IY8kSaqeTtVuYGVFxOsR8VJEzIqI2og4Ind9nYh4MCJeiYjnI2KXvHt6RMTtETEnIv4e\nEcPyahERV0XEq7l7T27w836Zq82JiAsb1I7P3TMnIkZHROeW/v6SJEmSJKlpeveG/faDkSNhyhT4\n4AN48EE46ywYPBg6N+G3+MWL4ckn4de/hj32gNVXh912g3PPzYKihQtb/GtIkiT9/9ripM4y4MiU\n0gsNro8Enkkp7R8RA4G7I6JvSmkpcAawMKW0eUT0BaZGxOMppQ+AY4D/SCltFhFrALNytZciYldg\nOLBN7uc+HRFPp5QejIh+wK+Br6eU3o2Ie4ETgD+0/H8CSZIkSZK0supCnv32y95//HHxJM+yZct/\nRl3IUxf0dO9ePMmzyiot+z0kSVLH1eYmdYDIrYaOBK4FSCnNAP4F7JarDc+rvQ48ARyad991udoH\nwFhgRF7tlpTSwpTSYuCGvNow4N6U0ru599fm1SRJkiRJUiu32mqw//5wySUwdWo2yfPAA/DTn8Kg\nQdCpCf/XZNEimDgRzj8fhg7NJnl23x3OOy+77iSPJEkqp7Y4qQNwS0QATAN+BiSgS0rp33mfeQPY\nOPd649z7Oq+voDYkr/ZUg9rwJjxTkiRJkiS1MXUhz/77Z+8//hgmTaqf5Jk5c8WTPHUhT13Q0707\n7LBD/STPkCFO8kiSpNK1xVBnl5TS/+bOr7kI+BPwHRqf3mlVTj/9dPr06VNwbcSIEYwY4YCPJEmS\nJEmtzWqrwQEHZAvgo48KQ57a2qaFPHWfhyzQaRjydO/eUt9AkiS1BmPGjGHMmDEF1z766KOSntXm\nQp2U0v/m/lwaEb8DXk4pzYuIJRGxbt60Tl9gbu71G8AmwDt5tYdzr+fmalMbua+uxhfUvvIFtUaN\nGjWKmpqaFX1FSZIkSZLUCvXpAwcemC0oLeRZuBAmTMgWGPJIktQRNDbcUVtby4ABA1b6WW3qTJ2I\n6BkR+aMuRwGzcq/vBE7KfW4QsAEwMVcbB5yYq/UjO2vnnlztLuAHEdEpItYk215tbF7tmIjoERHd\ngeOAO3K18cDBEbFuZHvBnZhXkyRJkiRJ7VxdyHPZZTB9Orz/PvzlL/CTn8CAAU07k6cu5Dn3XNht\nt+xMnj33hAsugKeeyiZ9JEmS6rS1SZ31gPER0Ylsu7V/km29BtnZOrdExCvAIuDbKaWludplwA0R\n8SqwBDg5pTQvV7sFGAjMAZYBl6eUZgOklCZGxFjgRbJze+5IKT2Qq70WEecCk3O1CcDolvvqkiRJ\nkiSpNVt9dTjooGwBfPhhNskzYUI2yTNrFqS0/GcsXAiPP54tyCZ5dtwxm+IZOhQGDXKSR5KkjizS\niv41oWaLiBpg5syZM91+TZIkSZKkDurDD7Ppm7rt2poS8jTUo0d9yLP77jB4MHTrVv5eJUlSy8rb\nfm1ASqm2qfe1tUkdSZIkSZKkNmn11eEb38gWwAcfFIY8zz674pDns8/gr3/NFtSHPEOHZiHPoEGG\nPJIktWeGOpIkSZIkSVWwxhpw8MHZgvKFPDvtVD/JY8gjSVL7YqgjSZIkSZLUCjQMeebNKwx5nnuu\naSHPY49lC6Bnz8KQZ+BAQx5JktoyQx1JkiRJkqRWaM014ZvfzBbUhzwTJtSHPCuyYAE8+mi2oDjk\nGTQIunZtoS8gSZLKzlBHkiRJkiSpDWgY8rz/fvEkz4o0FvLsvHPhJI8hjyRJrZehjiRJkiRJUhu0\n1lpwyCHZgvqQp26S5/nnV/yMBQvgkUeyBdCrV/0kz9ChMGCAIY8kSa2JoY4kSZIkSVI70FjI8+ST\n9ZM8TQl55s8vDnnyJ3kMeSRJqi5DHUmSJEmSpHZorbXg0EOzBfDee4UhzwsvrPgZ8+fDww9nC2DV\nVQtDnpoaQx5JkirJUEeSJEmSJKkDWHttOOywbEFpIc+nn8JDD2ULikOeAQOgi/+3SZKkFuNfs5Ik\nSZIkSR1Qw5Dn3XcLQ54XX1zxMxoLeXbZpXCSx5BHkqTy8a9VSZIkSZIksc46MGxYtqA+5JkwIQt5\nZs9e8TM+/RQefDBbUBjyDB0K/fsb8kiS1Bz+NSpJkiRJkqQiDUOef/+7cJKnlJCnd+/CSR5DHkmS\nVo5/bUqSJEmSJGmF1l0XDj88W5CFPBMn1oc8f/vbip/xySfwwAPZgsKQZ+hQ+PrXDXkkSVoe/5qU\nJEmSJEnSSlt3XTjiiGwBvPNO4SRPKSHPaqsVTvJsvTX06NEy/UuS1BYZ6kiSJEmSJKnZ1luvOOTJ\nn+R56aUVP+Pjj+H++7NVZ511YJNNsrXxxsWv11wTIlriG0mS1PoY6kiSJEmSJKns1lsPjjwyWwBv\nv51N8kyYkIU8f/97057z7rvZmjGj8XqvXvUBT8M/N9kENtjALd0kSe2Hf6VJkiRJkiSpxa2/fnHI\nkz/J09SQp6H587MpoC+aBOrcGTbcsPEpn7rXvXqV9rMlSao0Qx1JkiRJkiRV3Prrw/Dh2YL6kKdu\nkufll8vzc5YuhblzszVpUuOfWWut5Yc+66zjFm+SpNbBUEeSJEmSJElV1zDk+fTTLIh5441sNXz9\nr3/BsmXl+dnvv5+tWbMar6+ySuOhT92fX/4ydO1anl4kSVoeQx1JkiRJkiS1OquuCl/9arYa8/nn\nWbBTF/Y0FgB99ll5elm4EF55JVuNicjO7vmi0GeTTaB37/L0Iknq2Ax1JEmSJEmS1OZ07Qp9+2ar\nMSll0zdfNOnz/9q792g56/re4+9vEnLjkgRCwi0JEEDxAoIo4sK6dNWjRY8FKZ4iFcQLaGu9gEvq\nbZ1ewCWVVq3nrAIVDreCSj219pQFXWJFvFFKuMlFDIQk5EIIJCEkkISd7/nj94wMk2eyL8zes2fv\n92ut35rZz/Ob3/OdZ2ZgZz77+f2WLoW1aztTS2YJmFasgJ//vL7PzJn1U7s17s+ZAxMmdKYeSdLY\nZagjSZIkSZKkMScCZs8u7bWvre+zefPOr/R57LGyJk8nrF9f2t131++fPLmEPO2meZs3D6ZM6Uwt\nkqTeZagjSZIkSZKkcWn6dHj5y0ur09cHK1e+OOhpDn+WLoVNmzpTy9atsHhxaXUiyrpDO1vbZ+bM\nztQiSRq9DHUkSZIkSZKkGhMnlitk5s2r358J69a1D32WLYPHH+9MLZmwalVpt91W32ePPerX82nc\n33dfp3iTpF5nqCNJkiRJkiQNQQTsuWdpRx1V3+fZZ8s0bu3W9lm+HLZt60w9Tz8Nv/pVaXV22QUO\nOKB96DNvHkyb1plaJEnDw1BHkiRJkiRJGibTpsGhh5ZWp6+vXM3TLvRZurSENZ2wbRssWVJaO3Pm\ntA995s8vAVZEZ+qRJA2eoY4kSZIkSZLUJRMnwn77lXbccfV9NmzYMehpDn9WrepcPWvWlHb77fX7\nd91156HPfvvBJL9xlKRh439iJUmSJEmSpFFsxgw44ojS6mzZ8sIUb+3W9tm6tTO1bNoE999fWp2J\nE8sUb+3W9pk/vwRDkqShMdSRJEmSJEmSetiUKbBwYWl1tm8vV9/sLPRZt64ztfT1vTDurbfW95k9\nu33os2BB2e8Ub5JUz1BHkiRJkiRJGsMmTIB99int2GPr+2zcWL+eT+P+ypUlHOqEtWtLW7Sofv+0\naTsPffbfH3bZpTO1SFKvMdSRJEmSJEmSxrndd4dXvrK0Otu2wYoV7UOfpUvhuec6U8uzz8Kvf11a\nnQkTyto9rev5NIc/u+/emVokabQx1JEkSZIkSZK0U7vsAgceWFqdzHL1TWvQ0xz+PPlkZ2rZvr2s\nIfTYY/Czn9X3mTVr56HP3LlO8SapNxnqSJIkSZIkSXpJImDvvUs75pj6Pps27Rj2NIc+K1aUNXk6\nYd260u66q37/lCkwb96OU7s17s+bB5Mnd6YWSeokQx1JkiRJkiRJw27XXeHww0ur8/zzZe2edqHP\n0qWweXNnatmyBRYvLq1ORFmDqF3os2ABzJjRmVokaTAMdSRJkiRJkiR13aRJJTCZP79+fyY89VT7\n0GfZMlizpjO1ZMKqVaX98pf1ffbYY8egpzn82Wefsv6PJHWSoY4kSZIkSZKkUS8C9tqrtKOPru/z\n7LOwfPmOa/s07i9fXq4I6oSnn4Z77y2tzi67vDDFW13oM28eTJ3amVokjR+GOpIkSZIkSZLGhGnT\n4LDDSqvT1werV7cPfZYuhY0bO1PLtm3wyCOltTN3bv3Ubo37s2aVMEuSGgx1JEmSJEmSJI0LEyfC\n/vuX9sY37rg/EzZs2DHoaQ5/Vq/uXD2PP17a7bfX799tt/rQp3G7337lOUkaPwx1JEmSJEmSJIly\nVczMmaUdeWR9ny1bXpjirS70Wb4ctm7tTD3PPAP3319anYkT4YAD2oc+8+fD9OmdqUXS6GCoI2nM\nuu666zj11FO7XYY0bviZk0aenztpZPmZk0aenzuNRlOmwCGHlFZn+/Zy9U270GfZMli/vjO19PW9\nMG47s2e3D30WLChrFDWmePMzJ41+hjqSxix/EZFGlp85aeT5uZNGlp85aeT5uVMvmjAB9t23tDe8\nob7P00/Xr+fTuL9yZZkKrhPWri3tjjvq90+fXkKe+fPhvvuu44YbTmXChPI8Jk588e1QtnVqnOEc\nu7/9rmuk0cRQR5IkSZIkSZJG0B57wKteVVqdrVthxYr2oc+yZfDcc52pZfNmePDB0gCuuaYz444l\nEaM/eBoNY4+GccZDAGeoI0mSJEmSJEmjyOTJcNBBpdXJhCee2DH0aQ5/nnpqZGseyzLLVHd9fbBt\nW7erUX9Ge/DUuP/kk0N7foY6kiRJkiRJktRDImDOnNJe97r6Ps88s2PY0xz6rFhR1v+RxpqxHsAZ\n6oyMqQAPPPBAt+uQxpUNGzawaNGibpchjRt+5qSR5+dOGll+5qSR5+dOeunmzi2tNfx5/vlytc+q\nVbB6dbn913/dwFvfuojt23lRa1ypkll+btxv3tbfY3bWmvsO5THt9kuj22/zgqmDeVSk7+5hFxHv\nA/6x23VIkiRJkiRJkqRR5bTMvHagnQ11RkBE7AW8HXgU6NASZpIkSZIkSZIkqUdNBQ4EbsrMAa+w\nY6gjSZIkSZIkSZLUAyZ0uwBJkiRJkiRJkiT1z1BHkiRJkiRJkiSpBxjqSJIkSZIkSZIk9QBDnREQ\nEYdExM8i4tcRcVtEHN7tmqSxKiK+ERFLImJ7RBzR7Xqk8SAipkTEP0fEgxFxZ0TcFBELu12XNJZV\nn7O7qs/cLRHxmm7XJI0HEXFm9Xvmu7tdizTWRcSjEfFA9f+6RRFxSrdrksayiJgcEd+MiIci4u6I\nuKrbNUljWUTs2fT/uEVVdrA1Imb299hJI1GguAS4ODOvjoiTgSuB13e5Jmmsuh64EPhptwuRxplL\nMvNGgIj4E+BbwFu6W5I0pp2SmU8DRMSJwBWAwY40jCJiAfBh4BfdrkUaJ7YD783Me7tdiDROXAhs\nz8zDACJiTpfrkca0zHwKOKrxc0ScC/xOZq7v77FeqTPMImJv4LXAPwJk5veAeRFxcFcLk8aozPxp\nZq4Eotu1SONFZm5pBDqVXwILulWPNB40Ap3KTMoXX5KGSUQE5Q8WPg5s7XI50ngR+O86aURExHTg\ng8AXGtsyc033KpLGpQ9Rft/sl6HO8JsHrMrM5n9oLwPmd6keSZKG2yeB73e7CGmsi4grI2IZ8BfA\n+7tdjzTGnQPcmpl3drsQaZy5upoG6h8iYna3i5HGsIXAU8AXIuL2anrft3a7KGm8iIg3Uv5Y798G\n0t9QR5IkdUxEfJ7yD4LPd7sWaazLzDMycz7wReCvu12PNFZFxCuBk4ELul2LNM68KTOPBI4GnqRM\nZS9peEyizLbwq8x8HeUP9b5TzUAkafh9ELiq5cKQtgx1ht9yYN+IaD7X8ylX60iSNGZExGeAE4F3\nZOZz3a5HGi8y82rgLRExq9u1SGPUmyhfdP0mIpYAbwAujYizu1uWNLZl5mPVbR/wdeD47lYkjWnL\ngD7gWoDMvAtYAry6m0VJ40FE7Aq8F7h8oI8x1BlmmfkEsIhqSoyI+ANgeWY+0tXCJEnqoIg4B/hD\n4G2ZubHb9UhjWUTMiIh9m34+EVibmeu6WJY0ZmXmxZm5f2YenJkHUdaOOyszL+l2bdJYFRHTI2JG\n06b3AU5/KA2TzHwSuBl4B0BEHAQcCDzQxbKk8eIPgbsy86GBPmDSMBajF3wUuKKakmYDcGaX65HG\nrIi4GHgnMBe4KSI2ZuZhXS5LGtMiYn/gIuBh4D+qxaSfy8zjuluZNGbNAK6PiKlAAmuAd3W3JGlc\nyW4XII0Dc4HvVbOeBPAIcHp3S5LGvI8Bl0XEhZSrds7KzFVdrkkaD84ELh3MAyLT30clSZIkSZIk\nSZJGO6dfkyRJkiRJkiRJ6gGGOpIkSZIkSZIkST3AUEeSJEmSJEmSJKkHGOpIkiRJkiRJkiT1AEMd\nSZIkSZIkSZKkHmCoI0mSJEmSJEmS1AMMdSRJkiRJkiRJknqAoY4kSZIkSZIkSVIPMNSRJEmS1FER\nMSci/iki1kZEX0R8YpiPd0VELBnG8T8QEdsjYv4wHmNBdYzTh+sYkiRJknqfoY4kSZI0ykXEGdUX\n/psjYt+a/T+OiHu6UVsbXwfeBlwAvB+4cZiPl8D2YR4/h3F8jRIR8bmI+P1u1yFJkiS1Y6gjSZIk\n9Y4pwJ/VbB9tgcNbgO9n5tcy89rMfGiYj/dh4OXDOP5VwLTMXDaMx9Do8HnAUEeSJEmjlqGOJEmS\n1DvuAj4SEft0u5B+zAE2DPdBImI6QGb2Zea24TpOFluHa3xJkiRJGihDHUmSJKk3JPBlYBL1V+u8\nSERMjIgvRcTiiHguIpZExAURMXmoBUTEQRFxfUQ8GRGbIuIXEXFC0/4zIqIxDdrHqynj+nYyXmMd\nmXMi4lMR8Wg1xdyPI+KVLX2viIiNEXFwRNwQEU8D1zTtW9Jm3I80nYP/jIhjaup4WUR8NyLWVMd/\nMCLOb9q/w5o6Va0/iIi3RcSdEfFsRNwXESe1jD0rIi6KiHuq+jdU9R8x8DO/Q70zIuJr1Wv6XEQs\nj4grI2LPpj57R8RlEbG6qu2u1vV6Ws7TH0fEw9XrelNE7F/1+VI1/uaI+H5EzGwZY0Dnoeq70/dP\n1efNVU2nRMQXqmM/GxE/jIiFNWMeGxE3RsT6aswfR8QbW/r8eTXmwuq9sq7qf3lETG3qtx2YDjRe\n7+0RcXm1b7eI+HrTOX88Iv49Il4zmNdOkiRJeqkmdbsASZIkSQO2hDIV2Eci4iuZuXonfS8DTge+\nC1wEHAt8jjJN2cmDPXBEzAF+AUwFvgE8BZwB/CAiTs7MfwFuAf6IErb8e1XrQJwB7Ab8r2r8TwI3\nR8SrM/OJqk9S/v1yE3ArcC6wuWlf3RR0p1XjXlztPw/4XkQcnJl91fM6ohpvC3AJsBRYCLwL+OJO\nxk/gMODb1fhXAGcC10fE2zPz5qrfwcC7gespr99c4GzgxxHxin5ewx1ExK7AT4GXUV7jO4HZ1TEO\nAJ6qgopbqmN/E3gUOAW4IiJmZOY3W4b9I2AX4O+APavzdH1E/Ah4M/AV4BDgE5T30ocHex4G+P5p\n9mdAH/BVYEZV0zXAcU3n4q3ADcB/AX9OWVfpTOBHEXF8Zv5XU41QPguPVGMfXT2Pxymfi8Z5uAy4\nDbi02vZwdXsJ8J7qfD4A7AUcDxxOuYJOkiRJGhGGOpIkSVJvuYAS1pwHfLquQxVUnA5cmpkfrTZf\nHBFPAOdGxJsz85ZBHvdzwN7A8Zn5i+o43wLuAf4W+JfMfBR4NCKuAR7KzGsHOPZC4JBrGn5SAAAH\nCUlEQVRGwBERN1G+WD8P+ExTv8nAdzLzizsOUWteNe7T1bgPAd8H3k4JA6B8SZ/AUZm5ouX59udQ\n4D2NQKK6quNB4EKgcUXQPZl5WPODIuJq4NfAhyiv52B8FngFcFJm/qBp+5eb7p9NCX1Oy8xvV8e8\nGPgJcH5EXJ6Zm5r670c5T89UfSdRnv9U4JjM3F5tnwOcFhEfa5nubiDnod/3T8vznAIc2RS+rQe+\nXgVh91d9/h64OTPf2XhQRFwC3A+cD7yjZcw7MvOspr6zKa/B5wAy89rq8Y/UvHdPAP4hMz/btO0i\nJEmSpBHm9GuSJElSD8nMJcDVwFkRMbdNtxMoQcXXWrb/DRDAO3d4RP9+D/jPxhfyVS2bKFc0HBgR\nrxjCmA3/3HzFSmbeTgl1Tqjpe/Egxv12I9Cp3Ep5/gfDb7/UfxNwWUugM1Arm68wycyNlKuTjqoC\nEJrDj4iYUE2RtpkS6hw9hGO+B7i7JdBp9XvA6kagU9XRR7kSZzfK1TfNvtsIdCq3VbdXNwKdpu2T\ngf1bHt/veWDw75/LG4FOpfW1ew0lTLouIvZqNGB34Gbgd1rGS8rVNs1uBfaKiN3o33rg2IjYdwB9\nJUmSpGFjqCNJkiT1nvMp02W1W1tnAWUqqsXNGzPzccqX0wuGcMwFlCCi1QNN+4dqcc22h4ADW7Y9\nn5mPDWLc5c0/ZOb66u6s6vbg6va+QYzZrF3dUNUexaerq4S2AGuBNcCrKdOKDdZC4Ff99FkA/KZm\n+wOUYKT1tVre8vOG6rb1XDe2z2rZ3u95YPDvn9aa1rUc+9Dq9irgiaa2hjKt2uSIaD2/y/oZc2c+\nC7wKWB4Rt0XE/4yIgwbwOEmSJKmjnH5NkiRJ6jGZuaSa4uysiLhwZ11HqqYRsmWQ/fvabI+XWsgg\nfAH4S+BblDV6nqIEbt9g9PyRXbvz1M3z19+xG+fuXODuNn2fafl5yM8nM6+PiJ8AJwH/jTIt4HkR\ncVJm3tTf4yVJkqROMdSRJEmSetP5lIXdz6vZt5TypfehNF0dUU2FNbPaP1hLKeu0tDq8af9QHVqz\n7TDg0Zcw5kA8Ut2+aoiPP6RmW+McPVrdngz8qHktF4CImEm5smSwHqb/epdSrgRq1YnXqs5AzkOn\n3z8PV7cbM/NHg3zszrQNQqsr3S6mrE81G7iTEtoZ6kiSJGnEjJa/DJMkSZI0CJn5CHANcDawT8vu\nGyhXH3yqZfu5lC+t/62xISL2iYiXRcTEfg55A/D6iDi26bG7AmcBS5oWrx+KEyNiv6ZxXw8cWx1z\n2GTmWuAnwAcjYt4QhtgvIk5q/BARewDvB+7MzDXV5j5argSJiFPYcV2agfoecGRE/P5O+twA7BMR\n/6PpmBOBPwU2ArcM8djtDOQ8dPr9cwcl2PlMNc6LVKHLUGyiBJ/NY02ontNvVe+dlcCUIR5HkiRJ\nGhKv1JEkSZJ6Q90UURdQvjx/GU3rrGTmPRFxJWV6tlmUL/GPBU4H/m9mNn+p/5Vq+4HsuOYILf1O\nBW6MiL+jTCP2AcpaKO8Z2lP6rcXATyPi74GpwCcpV7F89SWOOxCfAG4FFkXEpcAS4CDghMw8qp/H\nPgR8KyJeBzwOfAiYA5zR1Of/AV+KiMuBn1OuoDmNF640GayvAn8AXB8R/4cSbuwF/Hfg7My8F7iU\nEvZdERHHUK6WOQU4DvhkZm4a4rGh/n04kPPQ0fdPZmZEfJgSFt1XnYsVlLDsLZT1f3YWfLVzB/C7\nEfFpSmizhHK122MR8U+Uqd6eAd4GHAOcM4RjSJIkSUNmqCNJkiT1hh2mhcrMhyPiasqX5637P0QJ\nDj4AnAispoRAf1kz7vZ+D565JiKOAy4EPk4JX+4B3pWZN9aMOZj1fK6qavgUJQy4DfjTarqr1nHb\nljjAGl60vQrA3gD8FfBRyvNaCnxnAHX/hnL1y0WU6eKWAO/NzB829fkyMB14H/BeSmhwAiXkqKt5\npzJzU0QcD/wFZX2X04E1wA+Bx6o+z0XEm3khsNuDEkx8IDOvrjlmu/NUW0LNtn7PwxDeP/0eOzNv\nqcb8EvAnwG6U9/ltwCVtxujPOdVj/wqYBlxJuZrof1PW0jmJMuPFYuBjmXnpEI8jSZIkDUlkjrW1\nUyVJkiT1gohYQAkAPpOZf9vtegYjIpYA92bmu7tdSzd5HiRJkqSR5Zo6kiRJkiRJkiRJPcBQR5Ik\nSZIkSZIkqQcY6kiSJEnqpsGuvzNa9GrdneZ5kCRJkkaQa+pIkiRJkiRJkiT1AK/UkSRJkiRJkiRJ\n6gGGOpIkSZIkSZIkST3AUEeSJEmSJEmSJKkHGOpIkiRJkiRJkiT1AEMdSZIkSZIkSZKkHmCoI0mS\nJEmSJEmS1AMMdSRJkiRJkiRJknqAoY4kSZIkSZIkSVIP+P9+/uk2YAPAagAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a004a58ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(1, figsize=(20, 5))\n",
    "plt.plot(pca.explained_variance_, linewidth=3)\n",
    "plt.axis('tight')\n",
    "plt.title('Scree plot for PCA with FEATURE RE-ENGINEERED data')\n",
    "plt.xlabel('No. of principal components')\n",
    "plt.ylabel('Explained variance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: Elbow is used to choose the number of principal components (Here, #PCs chosen = 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.26567194  0.22292369  0.20453226  0.17938456]\n",
      "Cumulative explained variance:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ 26.57,  48.86,  69.31,  87.25])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Delay_YesNo = data['ArrDel15']\n",
    "#data.drop(['ArrDel15'], axis=1, inplace=True)#Removing target variable\n",
    "#print('Dimension reduced to:')\n",
    "#print(len(data.columns))\n",
    "#data.describe()\n",
    "pca = PCA(n_components=4)\n",
    "dataNotCollinear_reduced = pca.fit_transform(dataNotCollinear)\n",
    "print(pca.explained_variance_ratio_)\n",
    "print('Cumulative explained variance:')\n",
    "np.cumsum(np.round(pca.explained_variance_ratio_, decimals=4)*100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model building with dimension reduced data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Train/test split\n",
    "Train:Test split = 80:20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train, x_test, y_train, y_test = train_test_split(dataNotCollinear_reduced, Delay_YesNo, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Grid search\n",
    "Grid search is not performed for all the desired parameter combinations and values due to computational time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random forest Grid Search with dimension reduced data took [ 0:01:40.813686 ] seconds.\n"
     ]
    }
   ],
   "source": [
    "startTimeGS = datetime.now()\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "rf = RandomForestClassifier()\n",
    "param_grid = {\n",
    "                 'n_estimators': [10],\n",
    "#                 'min_samples_split': [2, 4],\n",
    "#                 'min_samples_leaf': [2, 4],\n",
    "#                 'max_features': [sqrt', 'log2'],\n",
    "                 \"criterion\" : [\"gini\"]\n",
    "             }\n",
    "grid_rf = GridSearchCV(rf, param_grid, cv=3)\n",
    "grid_rf.fit(x_train, y_train)\n",
    "bestModel = grid_rf.best_estimator_\n",
    "bestParameters = grid_rf.best_params_\n",
    "gridScores = grid_rf.grid_scores_\n",
    "print('Random forest Grid Search with dimension reduced data took [', datetime.now() - startTimeGS, '] seconds.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
      "            min_impurity_split=1e-07, min_samples_leaf=1,\n",
      "            min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
      "            n_estimators=10, n_jobs=1, oob_score=False, random_state=None,\n",
      "            verbose=0, warm_start=False)\n",
      "{'criterion': 'gini', 'n_estimators': 10}\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[mean: 0.60450, std: 0.00063, params: {'criterion': 'gini', 'n_estimators': 10}]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(bestModel)\n",
    "print(bestParameters)\n",
    "gridScores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Model selection\n",
    "- Best model and parameters from above are used to train the final model on entire training set\n",
    "- 3- fold Cross validation is performed to find the overall error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean cross validation score is: 0.604122234433\n",
      "Random forest training and testing with PCs took [ 0:01:45.095981 ] seconds.\n"
     ]
    }
   ],
   "source": [
    "startTimeRF = datetime.now()\n",
    "rf = RandomForestClassifier(n_estimators = bestParameters.get('n_estimators'), \n",
    "#                            min_samples_split=bestParameters.get('min_samples_split'),\n",
    "#                           min_samples_leaf = bestParameters.get('min_samples_leaf'),\n",
    "#                           max_features = bestParameters.get('max_features'),\n",
    "                           criterion = bestParameters.get('criterion'))\n",
    "cv = cross_validation.KFold(len(x_train), n_folds=3, shuffle=True, random_state=2)\n",
    "cvScores = cross_val_score(rf, x_train, y_train, cv=cv)\n",
    "print ('Mean cross validation score is: ' + str(np.mean(cvScores)))\n",
    "rf.fit(x_train, y_train)\n",
    "print('Random forest training and testing with PCs took [', datetime.now() - startTimeRF, '] seconds.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Performance evaluation\n",
    "- Prediction is done on the hold out test set to evaluate performance\n",
    "- Confusion matrix, Accuracy and Recall are computed\n",
    "- ROC curve is also plotted to pictorically depict 'Area under curve (AUC)' as 'Accuracy' "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 62.5%\n",
      "Recall: 55.47%\n",
      "Confusion matrix:\n",
      "[[50912 22287]\n",
      " [32756 40811]]\n",
      "Area under the ROC curve: 0.374862679361\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABoEAAAELCAYAAAAWSS8ZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJEmSJEmSJEkj6P8BaKDmzOxrOk0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a004433828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_rf_pred = rf.predict(x_test)\n",
    "\n",
    "labels = [0, 1]\n",
    "cm = confusion_matrix(y_test, y_rf_pred,labels)\n",
    "\n",
    "print('Accuracy: ' + str(np.round(100*float(cm[0][0]+cm[1][1])/float((cm[0][0]+cm[1][1] + cm[1][0] + cm[0][1])),2))+'%')\n",
    "print('Recall: ' + str(np.round(100*float((cm[1][1]))/float((cm[1][0]+cm[1][1])),2))+'%')\n",
    "print('Confusion matrix:')\n",
    "print(cm)\n",
    "\n",
    "fpr, tpr, _ = roc_curve(y_test, y_rf_pred)\n",
    "auc = np.trapz(fpr,tpr)\n",
    "print('Area under the ROC curve: ' + str(auc))\n",
    "\n",
    "fig = plt.figure(figsize=(20, 5))\n",
    "\n",
    "plt.subplot2grid((2, 2), (1, 0), colspan=1)\n",
    "#plt.subplot(1,2,1)\n",
    "plt.plot(fpr,tpr,color='green')\n",
    "plt.xlabel('False positive rate (FPR)')\n",
    "plt.ylabel('True positive rate (TPR)')\n",
    "plt.title('Receiver operating characteristic (ROC)')\n",
    "\n",
    "ax1 = plt.subplot2grid((2, 2), (1, 1))\n",
    "#ax = fig.add_subplot(122)\n",
    "cax = ax1.matshow(cm)\n",
    "plt.title('Confusion matrix for Random Forest classifier with PCs')\n",
    "fig.colorbar(cax)\n",
    "ax1.set_xticklabels([''] + labels)\n",
    "ax1.set_yticklabels([''] + labels)\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('True')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inferences 3:\n",
    "- Random forest on PCs had poorer performance by > 30%.\n",
    "- As good results have already been observed, it may be deduced that good predictors do exist in the dataset. But, its effect might have been blurred by other highly correlated variables thereby adding noise to the 'good' predictors.\n",
    "- PCA helps when we rotate axis in the direction of maximal variance. It might be that the rotation hurted here by mixing noisy dimensions with good predictors. This may be geometrically verified by projecting data back on to the components.   \n",
    "- PCA assumes linearity in data and therefore it was a bad idea to do dimension reduction with PCA, given that the data is non-linear. I envisage that the data potentially consists of multiple separate clusters.  \n",
    "- Thus tailoring features too closely to the assumptions of a different model (e.g. Linear Regression) caused the performance to deteriorate with Random Forest\n",
    "- It is also inferred that dimension reduction with PCA is not needed for Random forest as it is already a modeling method that inherently handles variance in data through bagging, random splitting and voting. It is best to give all the parameters in the original form as seen in PART 2.\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Future Work:\n",
    "- As performance was observed to be improving with more data, it is strongly suggested to train on more data which is easily available from DOT website.\n",
    "- Morever as sampling is done to balance the classes, it may be verified that there is equal representation of positive and negative samples from each month.\n",
    "- These will improve Recall and give better prediction for 'Delayed' flights which is the problem in hand.\n",
    "- Sampling of data to avoid class imbalance resulted in discarding a lot of data. I am not much concerned about this due to availability of huge amounts of data in DOT website. But, this may need to be addressed elsewhere by exploring imputation, possibly mode imputation.\n",
    "- Feature creation could have been explored by deriving new features from existing data. Examples can be Airport delay, Airline delay etc from taxi duration, wheels on, wheels off etc. This might have provided much more intuitive analysis results.\n",
    "- Feature scaling is also an option that may be considered as it is likely to enhance performance.\n",
    "- Grid search with more trees and other parameters in Random Forest are essential and is highly likely to give improved results due to more stability.\n",
    "- It is also recommended that an initial coarser search followed by finer grid search can help in identifying optimal model parameters that can enable prediction with higher performance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### References:\n",
    "- Kaggle.com\n",
    "- stats.stackexchange.com\n",
    "- fivethirtyeight.com"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overall execution took [ 0:11:10.341817 ] seconds.\n"
     ]
    }
   ],
   "source": [
    "print('Overall execution took [', datetime.now() - startTime, '] seconds.')"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}