Step 2a-Approach1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cross_validation.py:41: 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": [
    "##library  dependencies\n",
    "from sklearn import neighbors\n",
    "from sklearn.model_selection import train_test_split\n",
    "from numpy import loadtxt\n",
    "from numpy import sort\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import classification_report\n",
    "from xgboost import XGBRegressor\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.metrics import f1_score\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "import bokeh\n",
    "import sklearn\n",
    "import xgboost as xgb\n",
    "import numpy as np\n",
    "from scipy.stats import norm\n",
    "import pandas as pd\n",
    "import matplotlib.mlab as mlab\n",
    "import sqlite3\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy import stats\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Approach 1\n",
    "\n",
    "-  In this approach the properties data (27 million records) and transaction data (90k records) are merged together before imputing any missing values "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import Data\n",
    "-  Importing properties_2016 which has data about 27 million unique house properties with 58 attributes\n",
    "\n",
    "-  Importing transaction data which has 90k records of the properties sold in the year 2016\n",
    "\n",
    "-  Merging the data sets based on property_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:2698: DtypeWarning: Columns (22,32,34,49,55) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  interactivity=interactivity, compiler=compiler, result=result)\n"
     ]
    }
   ],
   "source": [
    "##import properties 2016  data\n",
    "prop2016= pd.read_csv('C:\\\\Users\\\\Dell\\\\Desktop\\\\inf552\\\\Zillow Prize\\\\properties_2016.csv')\n",
    "##import train 2016  data\n",
    "train2016=pd.read_csv('C:\\\\Users\\\\Dell\\\\Desktop\\\\inf552\\\\Zillow Prize\\\\train_2016.csv',parse_dates=[\"transactiondate\"])\n",
    "##merge\n",
    "df_train = train2016.merge(prop2016, how='left', on='parcelid')\n",
    "df_train['transaction_month'] = df_train['transactiondate'].dt.month"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Renaming attributes names to be meaningful"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train=df_train.rename(index=str, columns={\"pooltypeid7\": \"pool_with_spa_tub_no\", \n",
    "                                             \"pooltypeid2\": \"pool_with_spa_tub_yes\"})\n",
    "\n",
    "df_train=df_train.rename(index=str, columns={\"hashottuborspa\": \"has_hottub_or_spa\", \n",
    "                                             \"pooltypeid10\": \"just_hottub_or_spa\"})\n",
    "\n",
    "df_train=df_train.rename(index=str, \n",
    "                         columns={\"calculatedfinishedsquarefeet\": \"total_finished_living_area_sqft\", \n",
    "                                  \"finishedsquarefeet12\": \"finished_living_area_sqft\",\n",
    "                                  \"lotsizesquarefeet\": \"lot_area_sqft\",\n",
    "                                  \"finishedsquarefeet50\": \"finished_living_area_entryfloor_sqft1\",\n",
    "                                  \"finishedfloor1squarefeet\": \"finished_living_area_entryfloor_sqft2\",\n",
    "                                  \"finishedsquarefeet6\": \"base_unfinished_and_finished_area_sqft\",\n",
    "                                  \"finishedsquarefeet15\": \"total_area_sqft\",\n",
    "                                  \"finishedsquarefeet13\": \"preimeter_living_area_sqft\",\n",
    "                                  \"taxvaluedollarcnt\":\"total_parcel_tax\",\n",
    "                                  \"landtaxvaluedollarcnt\":\"land_tax\",\n",
    "                                  \"taxamount\":\"total_property_tax_2016\",\n",
    "                                  \"structuretaxvaluedollarcnt\":\"structure_tax\",\n",
    "                                  \"garagetotalsqft\":\"garage_sqft\",\n",
    "                                  \"fireplacecnt\":\"fireplace_count\",\n",
    "                                  \"buildingqualitytypeid \":\"building_quality_id\",\n",
    "                                  \"heatingorsystemtypeid\":\"heating_system_id\",\n",
    "                                  \"airconditioningtypeid\":\"ac_id\"\n",
    "                                 \n",
    "                                 \n",
    "                                 \n",
    "                                 })\n",
    "df_train=df_train.rename(index=str, columns={\"decktypeid\": \"deck_flag\"})\n",
    "\n",
    "df_train=df_train.rename(index=str, columns={\"yardbuildingsqft17\": \"patio_sqft\",\n",
    "                                             \"yardbuildingsqft26\":\"storage_sqft\"})\n",
    "\n",
    "df_train=df_train.rename(index=str, columns={\"bathroomcnt\": \"total_bath\",\n",
    "                                             \"fullbathcnt\": \"full_bath\",\n",
    "                                             \"threequarterbathnbr\": \"half_bath\"})\n",
    "\n",
    "df_train=df_train.rename(index=str, columns={\"storytypeid\": \"basement_flag\"})\n",
    "\n",
    "df_train=df_train.rename(index=str, columns={\"poolsizesum\": \"pool_sqft\"})\n",
    "\n",
    "             "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dealing Attributes with Missing Values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  Pool_count is a binary variable, hence replace all NULL values with zero"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.loc[df_train.poolcnt.isnull(),'poolcnt']= 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  pool_with_spa_tub_no & pool_with_spa_tub_yes are again binary variables hence replace all NULL values with zero"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.loc[df_train.pool_with_spa_tub_no.isnull(),'pool_with_spa_tub_no']=0\n",
    "df_train.loc[df_train.pool_with_spa_tub_yes.isnull(),'pool_with_spa_tub_yes']=0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fixing contradictions in pool related variables\n",
    "\n",
    "-  When pool is present and if it has tub/spa then just_hottub_or_spa =0\n",
    "\n",
    "-  When there is no pool and if there is tub/spa then just_hottub_or_spa =1\n",
    "\n",
    "-  As they are binary variables convert NaN's to Zero"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "#when poolcnt=1 & has_hottub_or_spa=1 & just_hottub_or_spa is null then just_hottub_or_spa =0\n",
    "#when poolcnt=0, has_hottub_or_spa=1, just_hottub_or_spa =1\n",
    "\n",
    "df_train.loc[ (df_train.poolcnt==1) & (df_train.has_hottub_or_spa==1) & (df_train.just_hottub_or_spa.isnull()),'just_hottub_or_spa']=0\n",
    "             \n",
    "#has_hottub_or_spa is null and just_hottub_or_spa is null, both has to be zero\n",
    "\n",
    "df_train.loc[ (df_train.has_hottub_or_spa.isnull()) & (df_train.just_hottub_or_spa.isnull()),['has_hottub_or_spa','just_hottub_or_spa']]=0\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  When there is no pool, make poolsize as zero instead of Nan\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "16932\n"
     ]
    }
   ],
   "source": [
    "df_train.loc[ df_train.poolcnt==0,'pool_sqft']=0\n",
    "print(df_train.pool_sqft.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  basement_flag has values 7 & Null hence convert it to a binary variable with value of zero and 1\n",
    "-  When basement_flag is zero make basement_sqft also zero\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.loc[df_train.basement_flag.isnull(),'basementsqft']=0\n",
    "df_train.loc[df_train.basement_flag.isnull(),'basement_flag']=0\n",
    "df_train.loc[df_train.basement_flag==7,'basement_flag']=1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  There seems to be inconsistency between the fireplace_flag and fireplace count, Let's fix it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "after 222\n"
     ]
    }
   ],
   "source": [
    "#df_train.fireplaceflag.isnull().sum()\n",
    "#df_train.fireplace_count.isnull().sum()\n",
    "df_train.loc[(df_train.fireplace_count.isnull()) & (df_train.fireplaceflag.isnull()),'fireplaceflag'] = False\n",
    "df_train.loc[(df_train.fireplace_count.isnull()) & (df_train.fireplaceflag==False),'fireplace_count'] = 0\n",
    "df_train.loc[df_train['fireplace_count']>0,'fireplaceflag']= True\n",
    "print(\"after\",df_train.fireplace_count.isnull().sum())\n",
    "#print(\"after\",df_train.fireplace_count.value_counts())\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  Dropping transaction date column as this doesn't have any corellation with target variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train=df_train.drop('transactiondate',axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  Garage count and Garage size have same number of missing values. Let's assume this is because when there are properties with no garages then both variables are NA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "after 0\n",
      "after 8920\n"
     ]
    }
   ],
   "source": [
    "df_train.loc[df_train.garage_sqft.isnull() & df_train.garagecarcnt.isnull(),['garagecarcnt','garage_sqft']]=0\n",
    "df_train.loc[(df_train.garagecarcnt>0) & (df_train.garage_sqft==0),'garage_sqft']=np.nan\n",
    "print(\"after\",df_train.garagecarcnt.isnull().sum())\n",
    "#print(\"after\",df_train.garagecarcnt.value_counts())\n",
    "print(\"after\",df_train.garage_sqft.isnull().sum())\n",
    "#print(\"after\",df_train.garage_sqft.value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  total_bath & calculatedbathnbr are duplicates , and calculatedbathnbr has more nulls, hence we will drop it\n",
    "-  if full_bath is null and half_bath is also null, let's make total_bath=0 (missing values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1165\n",
      "1182\n",
      "1182\n"
     ]
    }
   ],
   "source": [
    "#total_bath & calculatedbathnbr are duplicates , and calculatedbathnbr has more nulls, hence drop it\n",
    "\n",
    "df_train=df_train.drop('calculatedbathnbr',axis=1)\n",
    "\n",
    "# full_bath is null & half_bath is null & total_bath=0 (missing values)\n",
    "df_train.loc[(df_train.full_bath.isnull()) & (df_train.half_bath.isnull()) & (df_train.total_bath==0),'total_bath']=np.nan\n",
    "\n",
    "\n",
    "# when full_bath=total_bath, half_bath=0 \n",
    "\n",
    "df_train.loc[(df_train.full_bath==df_train.total_bath) ,'half_bath']=0\n",
    "\n",
    "# when total_bath is present but full and half bath is null\n",
    "# all 3 are null somemtimes\n",
    "\n",
    "print(df_train.total_bath.isnull().sum())\n",
    "print(df_train.half_bath.isnull().sum())\n",
    "print(df_train.full_bath.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  Assuming if these patio and shed variables has null values then there is no shed or patio in yard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#yardbuildingsqft17-patio in yard\n",
    "#yardbuildingsqft26- storage shed in yard\n",
    "df_train.loc[df_train.patio_sqft.isnull() ,'patio_sqft']=0\n",
    "df_train.loc[df_train.storage_sqft.isnull() ,'storage_sqft']=0\n",
    "print(df_train.patio_sqft.isnull().sum())\n",
    "print(df_train.storage_sqft.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### code fips code with respective county names\n",
    "    -  6037- LA\n",
    "    -  6059- Orange_County\n",
    "    -  6111- Ventura"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "LA               58574\n",
      "Orange_County    24505\n",
      "Ventura           7196\n",
      "Name: fips, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "df_train.loc[df_train.fips==6037 ,'fips']=\"LA\"\n",
    "df_train.loc[df_train.fips==6059 ,'fips']=\"Orange_County\"\n",
    "df_train.loc[df_train.fips==6111 ,'fips']=\"Ventura\"\n",
    "print(df_train.fips.isnull().sum())\n",
    "print(df_train.fips.value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### scaling down the latitude and longitide \n",
    "-  Knn imputation takes more time because of the hude numbers, moreover standardizing gives better results on most algorithms\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train['latitude']=df_train['latitude'].divide(100000)\n",
    "df_train['longitude']=df_train['longitude'].divide(100000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  deck_flag has only 2 values 66 or null- convert it into binary flag"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "df_train.loc[df_train.deck_flag==66 ,'deck_flag']=1\n",
    "df_train.loc[df_train.deck_flag.isnull() ,'deck_flag']=0\n",
    "print(df_train.deck_flag.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Imputing unit count based on property land type (Mode Imputation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "69705\n",
      "69700\n",
      "31922\n",
      "89976\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#numberofstories & unitcnt &roomcnt\n",
    "\n",
    "df_train.loc[df_train.roomcnt==0 ,'roomcnt']=np.nan\n",
    "\n",
    "print(df_train.numberofstories.isnull().sum())\n",
    "print(df_train.roomcnt.isnull().sum())\n",
    "print(df_train.unitcnt.isnull().sum())\n",
    "\n",
    "# propertylandusetypeid  and unitcnt is related \n",
    "\n",
    "#246 -Duplex (2 Units, Any Combination)\n",
    "#247 -Triplex (3 Units, Any Combination)\n",
    "#248 -Quadruplex (4 Units, Any Combination)\n",
    "#260 -Residential General\n",
    "#261 -Single Family Residential\n",
    "#263 -Mobile Home\n",
    "#264 -Townhouse\n",
    "#266 -Condominium\n",
    "#267 -Cooperative\n",
    "#269 -Planned Unit Development\n",
    "#275 -Residential Common Area \n",
    "#31 - Commercial/Office/Residential Mixed Used\n",
    "#47 -Store/Office (Mixed Use)\n",
    "#265 -Cluster Home\n",
    "\n",
    "df_train.loc[(df_train.propertylandusetypeid==31) & (df_train.unitcnt.isnull()),'unitcnt']=2\n",
    "df_train.loc[(df_train.propertylandusetypeid==47) & (df_train.unitcnt.isnull()),'unitcnt']=2\n",
    "df_train.loc[(df_train.propertylandusetypeid==246) & (df_train.unitcnt.isnull()),'unitcnt']=2\n",
    "df_train.loc[(df_train.propertylandusetypeid==247) & (df_train.unitcnt.isnull()),'unitcnt']=3\n",
    "df_train.loc[(df_train.propertylandusetypeid==248) & (df_train.unitcnt.isnull()),'unitcnt']=4\n",
    "df_train.loc[(df_train.propertylandusetypeid==260) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==261) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==263) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==264) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==266) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==267) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==269) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==275) & (df_train.unitcnt.isnull()),'unitcnt']=1\n",
    "\n",
    "#typeconstructiontypeid (based on location and year of building)\n",
    "print(df_train.typeconstructiontypeid.isnull().sum())\n",
    "print(df_train.propertylandusetypeid.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  \"preimeter_living_area_sqft\" and \"total_finished_living_area_sqft\" have the same values except that \"preimeter_living_area_sqft\" has more duplicates, Let's drop it\n",
    "\n",
    "-  \"total_area_sqft\" and \"total_finished_living_area_sqft\" have the same values except that \"total_area_sqft\" has more duplicates, Let's drop it\n",
    "\n",
    "-  \"total_finished_living_area_sqft\" and \"finished_living_area_sqft\" have the same values except that \"finished_living_area_sqft\" has more duplicates, Let's drop it\n",
    "\n",
    "-  \"base_unfinished_and_finished_area_sqft\" and \"total_finished_living_area_sqft\" have the same values except that \"base_unfinished_and_finished_area_sqft\" has more duplicates, Let's drop it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "df_train=df_train.drop('preimeter_living_area_sqft', axis=1)\n",
    "df_train=df_train.drop('total_area_sqft', axis=1)\n",
    "df_train=df_train.drop('finished_living_area_sqft', axis=1)\n",
    "df_train=df_train.drop('base_unfinished_and_finished_area_sqft', axis=1)\n",
    "\n",
    "#calculatedfinishedsquarefeet\": \"total_finished_living_area_sqft\", \n",
    "#\"finishedsquarefeet12\": \"finished_living_area_sqft\",\n",
    "#\"lotsizesquarefeet\": \"lot_area_sqft\",\n",
    "#\"finishedsquarefeet50\": \"finished_living_area_entryfloor_sqft1\",\n",
    "#finishedfloor1squarefeet\": \"finished_living_area_entryfloor_sqft2\",\n",
    "#\"finishedsquarefeet6\": \"base_unfinished_and_finished_area_sqft\",\n",
    "#\"finishedsquarefeet15\": \"total_area_sqft\",\n",
    "#\"finishedsquarefeet13\": \"preimeter_living_area_sqft\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  From the analysis different counties follow different land use code, comparing different counties, zillow has created it's own \"propertylandusetypeid\" matching all the counties, to reduce number of levels, hence we can drop \"propertycountylandusecode\"\n",
    "\n",
    "-  Same applies to propertyzoningdesc as well"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "#propertycountylandusecode\n",
    "#print(df_train.propertycountylandusecode.isnull().sum())\n",
    "\n",
    "df_train=df_train.drop(['propertycountylandusecode'], axis=1)\n",
    "\n",
    "#propertyzoningdesc, same reason as above applies to this column as well\n",
    "df_train=df_train.drop(['propertyzoningdesc'], axis=1)\n",
    "#print(df_train.propertyzoningdesc.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  Most zip id's either invalid or out of city, since enough information about location is given in latitude and longitude let's drop other location related fields\n",
    "\n",
    "    -  regionidcity\n",
    "    -  regionidzip\n",
    "    -  regionidneighborhood\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train=df_train.drop(['regionidzip','regionidcity','regionidneighborhood'], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  assessmentyear has a constant value for all the rows's, Let's drop it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.assessmentyear.value_counts()\n",
    "df_train=df_train.drop(['assessmentyear'], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  if Tax deliquency flag is null, it is safe to assume there is no unpaid tax for the property"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.loc[df_train.taxdelinquencyflag.isnull(),'taxdelinquencyflag'] = \"No\"\n",
    "df_train.loc[df_train.taxdelinquencyflag=='Y','taxdelinquencyflag'] = \"Yes\"\n",
    "df_train.loc[df_train.taxdelinquencyflag=='No','taxdelinquencyyear'] = 0\n",
    "\n",
    "#print(df_train.taxdelinquencyyear.value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  Convert yy to yyyy format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.loc[df_train.taxdelinquencyyear==99,'taxdelinquencyyear'] = 1999\n",
    "df_train.loc[df_train.taxdelinquencyyear==6,'taxdelinquencyyear'] = 2006\n",
    "df_train.loc[df_train.taxdelinquencyyear==7,'taxdelinquencyyear'] = 2007\n",
    "df_train.loc[df_train.taxdelinquencyyear==8,'taxdelinquencyyear'] = 2008\n",
    "df_train.loc[df_train.taxdelinquencyyear==9,'taxdelinquencyyear'] = 2009\n",
    "df_train.loc[df_train.taxdelinquencyyear==10,'taxdelinquencyyear'] = 2010\n",
    "df_train.loc[df_train.taxdelinquencyyear==11,'taxdelinquencyyear'] = 2011\n",
    "df_train.loc[df_train.taxdelinquencyyear==12,'taxdelinquencyyear'] = 2012\n",
    "df_train.loc[df_train.taxdelinquencyyear==13,'taxdelinquencyyear'] = 2013\n",
    "df_train.loc[df_train.taxdelinquencyyear==14,'taxdelinquencyyear'] = 2014\n",
    "df_train.loc[df_train.taxdelinquencyyear==15,'taxdelinquencyyear'] = 2015\n",
    "\n",
    "#print(df_train.taxdelinquencyyear.value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.loc[df_train.taxdelinquencyyear>0,'taxdelinquencyyear'] = 2016 - df_train['taxdelinquencyyear']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-  rawcensustractandblock has 3 fields concatened together, let's do string split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "# both are float columns now\n",
    "#rawcensustractandblock\n",
    "s_rawcensustractandblock=df_train.rawcensustractandblock.apply(lambda x: str(x))\n",
    "\n",
    "df_train['census_tractnumber']=s_rawcensustractandblock.str.slice(4,11)\n",
    "df_train['block_number']=s_rawcensustractandblock.str.slice(start=11)\n",
    "df_train['block_number']=df_train['block_number'].apply(lambda x: x[:4]+'.'+x[4:]+'0' )\n",
    "df_train['block_number']=df_train['block_number'].apply(lambda x: int(round(float(x),0)) )\n",
    "df_train['block_number']=df_train['block_number'].apply(lambda x: str(x).ljust(4,'0') )\n",
    "\n",
    "#droping censustractandblock since this is just a duplicate of rawcensustractandblock\n",
    "df_train=df_train.drop('censustractandblock', axis=1)\n",
    "\n",
    "# drooping rawcensustractandblock, since it's already stored as substrings in different column names\n",
    "df_train=df_train.drop('rawcensustractandblock', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### At this point most of the missing values are logically filled. Even now if some fields are missing more than 95% of values, then it doesn't make sense to try and impute them, because it will only add noise to the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['buildingclasstypeid', 'architecturalstyletypeid', 'typeconstructiontypeid']\n"
     ]
    }
   ],
   "source": [
    "missingvalues_prop = (df_train.isnull().sum()/len(df_train)).reset_index()\n",
    "missingvalues_prop.columns = ['field','percentage']\n",
    "missingvalues_prop = missingvalues_prop.sort_values(by = 'percentage', ascending = False)\n",
    "#print(missingvalues_prop)\n",
    "missingvaluescols = missingvalues_prop[missingvalues_prop['percentage'] > 0.95].field.tolist()\n",
    "\n",
    "print(missingvaluescols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = df_train.drop(['buildingclasstypeid', 'architecturalstyletypeid', 'typeconstructiontypeid'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "# time to deal with missing values\n",
    "# highly related propertylandusetypeid\n",
    "#  unitcnt            360\n",
    "# it seems most missing values from unitcnt are from propertylandusetypeid=265, which is Cluster Home\n",
    "# google search shows they are group of houses with shared walls \n",
    "#http://www.nytimes.com/1983/06/05/realestate/patio-homes-gain-in-cluster-complexes.html\n",
    "# year built also matches 1970-80's\n",
    "# says the cluster is anywhere between 5 to 25, asssume 10 as reassonable count \n",
    "\n",
    "df_train.loc[df_train.propertylandusetypeid==265,'unitcnt']=10\n",
    "print(df_train.propertylandusetypeid.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#pool related attribute , is related propertylandusetypeid, and only single or duplex houses have pools \n",
    "\n",
    "#pool_sqft  16932\n",
    "poolsizesum_mean = df_train.loc[df_train['poolcnt'] > 0, 'pool_sqft'].mean()\n",
    "\n",
    "df_train.loc[df_train.pool_sqft.isnull(),'pool_sqft']=poolsizesum_mean\n",
    "print(df_train.pool_sqft.isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n",
      "380\n",
      "1\n",
      "1\n"
     ]
    }
   ],
   "source": [
    "#total_parcel_tax\n",
    "#structure_tax\n",
    "#land_tax\n",
    "#total_property_tax_2016\n",
    "#2)recalculate total_parcel_tax =structure_tax + land_tax\n",
    "\n",
    "# total_parcel_tax =structure_tax + land_tax\n",
    "#->structure_tax=total_parcel_tax -land_tax\n",
    "\n",
    "df_train.loc[(df_train.total_parcel_tax>0) & (df_train.land_tax>0),'structure_tax']=df_train['total_parcel_tax']-df_train['land_tax']\n",
    "\n",
    "#structure_tax, i see a lot of structure tax is 0's, those must be NA's\n",
    "\n",
    "df_train.loc[df_train.structure_tax==0,'structure_tax']=np.nan\n",
    "\n",
    "print(df_train.total_property_tax_2016.isnull().sum())\n",
    "print(df_train.structure_tax.isnull().sum())\n",
    "print(df_train.total_parcel_tax.isnull().sum())\n",
    "print(df_train.land_tax.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(90275, 45)"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# regionidcounty is exact copy of fips code, dropping the dulicate column\n",
    "\n",
    "df_train = df_train.drop(['regionidcounty'], axis=1)\n",
    "df_train.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1421\n"
     ]
    }
   ],
   "source": [
    "#*******************************\n",
    "#bedroomcnt #1421 zero bed room houses ??, observed it's missing all other room count also missing\n",
    "\n",
    "df_train.loc[df_train.bedroomcnt==0,'bedroomcnt']= np.nan\n",
    "print(df_train.bedroomcnt.isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1165\n",
      "1182\n",
      "1182\n",
      "1421\n",
      "1416\n"
     ]
    }
   ],
   "source": [
    "# propertylandusetypeid & total living area\n",
    "#                              total_bath           1165\n",
    "#                              full_bath           1182\n",
    "#                              half_bath           1182\n",
    "#                                bedroomcnt      1421\n",
    "#                              roomcnt           1416\n",
    "\n",
    "\n",
    "# roomcnt=(full_bath+half_bath)+ bedroomcnt\n",
    "# total_bath=fullbath+ 0.5(half_bath)\n",
    "\n",
    "#caluculate full bath and half bath again from total bath as, it has few extra columns, (fixes 500 missing values in roomcnt )\n",
    "\n",
    "df_train.loc[(df_train.full_bath.notnull()) \n",
    "             & (df_train.half_bath.notnull()) \n",
    "             & (df_train.bedroomcnt.notnull()) \n",
    "             & (df_train.roomcnt.isnull()),['roomcnt']]=df_train.full_bath + df_train.half_bath + df_train.bedroomcnt\n",
    "\n",
    "\n",
    "# most bedroom count and roomcount null are in same place\n",
    "# all column null count 1133 all columns are null\n",
    "\n",
    "print(df_train.total_bath.isnull().sum())\n",
    "print(df_train.full_bath.isnull().sum())\n",
    "print(df_train.half_bath.isnull().sum())\n",
    "print(df_train.bedroomcnt.isnull().sum())\n",
    "print(df_train.roomcnt.isnull().sum())\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "#numberofstories\t69705\n",
    "df_train.loc[(df_train.propertylandusetypeid==246) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==246) & (df_train.numberofstories.isnull()),'numberofstories']=2\n",
    "df_train.loc[(df_train.propertylandusetypeid==247) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==247) & (df_train.numberofstories.isnull()),'numberofstories']=2\n",
    "df_train.loc[(df_train.propertylandusetypeid==248) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==248) & (df_train.numberofstories.isnull()),'numberofstories']=2\n",
    "df_train.loc[(df_train.propertylandusetypeid==260) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==260) & (df_train.numberofstories.isnull()),'numberofstories']=2\n",
    "df_train.loc[(df_train.propertylandusetypeid==261) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==261) & (df_train.numberofstories.isnull()),'numberofstories']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==263) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==263) & (df_train.numberofstories.isnull()),'numberofstories']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==266) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==266) & (df_train.numberofstories.isnull()),'numberofstories']=1\n",
    "df_train.loc[(df_train.propertylandusetypeid==269) & (df_train.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==269) & (df_train.numberofstories.isnull()),'numberofstories']=2\n",
    "prop2016.loc[(prop2016.propertylandusetypeid==275) & (prop2016.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==275) & (df_train.numberofstories.isnull()),'numberofstories']=1\n",
    "prop2016.loc[(prop2016.propertylandusetypeid==267) & (prop2016.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==267) & (df_train.numberofstories.isnull()),'numberofstories']=1\n",
    "#https://en.wikipedia.org/wiki/Townhouse , typical town house are usualyy large, and has atleast 6 rooms\n",
    "df_train.loc[(df_train.propertylandusetypeid==264) & (df_train.numberofstories.isnull()),'numberofstories']=2\n",
    "prop2016.loc[(prop2016.propertylandusetypeid==31) & (prop2016.numberofstories.notnull()),'numberofstories'].mode()\n",
    "df_train.loc[(df_train.propertylandusetypeid==31) & (df_train.numberofstories.isnull()),'numberofstories']=2\n",
    "print(df_train.numberofstories.isnull().sum()) \n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "#fireplace_count\n",
    "df_train.loc[(df_train.fireplaceflag==1) & (df_train.fireplace_count.notnull()),'fireplace_count'].mode()\n",
    "df_train.loc[(df_train.fireplaceflag==1) & (df_train.fireplace_count.isnull()),'fireplace_count']=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEhCAYAAACDefxEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x24b01ee4f28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "import seaborn as sns\n",
    "color = sns.color_palette()\n",
    "sns.set(style=\"darkgrid\")\n",
    "\n",
    "\n",
    "ax = sns.countplot(x=\"buildingqualitytypeid\", data=df_train)\n",
    "\n",
    "plt.xticks(rotation='vertical')\n",
    "plt.title(\"Frequency of Bathroom count\", fontsize=15)\n",
    "plt.show()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD3CAYAAADmBxSSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x24b03823080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(df_train.yearbuilt,df_train.buildingqualitytypeid , 'ro')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "#from ggplot import *\n",
    "\n",
    "#ggplot(aes(x='latitude', y='longitude', color='buildingqualitytypeid'), data=df_train) +  geom_point() +   scale_color_gradient(low = 'red', high = 'blue')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "#location seems to be related to building quality, (knnclassifier)\n",
    "\n",
    "def fillna_knn( df, base, target):\n",
    "    data_colnames = [ target ] + base\n",
    "    #print(\"data_colnames\",data_colnames)\n",
    "    missing_values_boolflag = df[target].isnull() #true for missing rows, false for columns with values\n",
    "    #print(\"miss\",missing_values_boolflag.head())\n",
    "    not_missing_boolflag = ~missing_values_boolflag \n",
    "    #print(\"not miss\",not_missing_boolflag.head())\n",
    "    number_of_missing_val = missing_values_boolflag.sum()\n",
    "    print(\"# of miss\",number_of_missing_val)\n",
    "    not_missing_rows = df.loc[ not_missing_boolflag, data_colnames ]\n",
    "    #print(not_missing_rows.head())\n",
    "    Y = not_missing_rows[target]\n",
    "    X = not_missing_rows[base]\n",
    "    X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.20, random_state=3192,stratify=Y)\n",
    "    metrics       = ['euclidean'] \n",
    "    weights       = ['distance'] \n",
    "    numNeighbors  = [5,10,15,20,25]\n",
    "    param_grid    = dict(metric=metrics,weights=weights,n_neighbors=numNeighbors)\n",
    "    cv            = StratifiedKFold(n_splits=3,random_state=3192,shuffle=False)\n",
    "    grid = GridSearchCV(neighbors.KNeighborsClassifier(n_jobs=-1),param_grid=param_grid,cv=cv,scoring='f1_weighted',refit=True,return_train_score=True,verbose=1,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "    grid.fit(X_train ,Y_train)\n",
    "    #print(\"grid.cv_results_\",grid.cv_results_)\n",
    "    print(\"grid.best_estimator_\",grid.best_estimator_)\n",
    "    print(\"grid.best_params_\",grid.best_params_)\n",
    "    print(\"grid.scorer_\",grid.scorer_)\n",
    "    #print(\"grid.n_splits_\",grid.n_splits_)\n",
    "    y_true, y_pred = Y_test, grid.predict(X_test)\n",
    "    \n",
    "    Z = grid.predict(df.loc[missing_values_boolflag, base])\n",
    "    #df.loc[ missing_values_boolflag, target ]  = Z\n",
    "    return Z\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32911\n",
      "(90275, 45)\n",
      "32911\n",
      "(90272, 45)\n",
      "# of miss 32911\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   33.4s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "           metric_params=None, n_jobs=-1, n_neighbors=10, p=2,\n",
      "           weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 10, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(f1_score, pos_label=None, average=weighted)\n",
      "predicted output shape (32911,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "\n",
    "print(df_train.buildingqualitytypeid.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "temp['buildingqualitytypeid']=temp['buildingqualitytypeid'].fillna(-1)\n",
    "temp=temp.groupby(\"buildingqualitytypeid\").filter(lambda x: x.buildingqualitytypeid.size > 3)\n",
    "temp['buildingqualitytypeid'] = temp['buildingqualitytypeid'].replace(-1,np.nan)\n",
    "print(temp.buildingqualitytypeid.isnull().sum())\n",
    "print(temp.shape)\n",
    "\n",
    "missing_values=fillna_knn(temp,\n",
    "                  base = [ 'latitude', 'longitude' ] ,\n",
    "                  target = 'buildingqualitytypeid')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['buildingqualitytypeid'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'buildingqualitytypeid' ]  = missing_values\n",
    "\n",
    "\n",
    "print(df_train.buildingqualitytypeid.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34195\n",
      "(90275, 45)\n",
      "34195\n",
      "(90269, 45)\n",
      "# of miss 34195\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   32.6s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "           metric_params=None, n_jobs=-1, n_neighbors=5, p=2,\n",
      "           weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 5, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(f1_score, pos_label=None, average=weighted)\n",
      "predicted output shape (34195,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "\n",
    "print(df_train.heating_system_id.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "temp['heating_system_id']=temp['heating_system_id'].fillna(-1)\n",
    "temp=temp.groupby(\"heating_system_id\").filter(lambda x: x.heating_system_id.size > 3)\n",
    "temp['heating_system_id'] = temp['heating_system_id'].replace(-1,np.nan)\n",
    "print(temp.heating_system_id.isnull().sum())\n",
    "print(temp.shape)\n",
    "\n",
    "missing_values=fillna_knn(temp,\n",
    "                  base = [ 'latitude', 'longitude' ] ,\n",
    "                  target = 'heating_system_id')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['heating_system_id'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'heating_system_id' ]  = missing_values\n",
    "\n",
    "\n",
    "print(df_train.heating_system_id.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "#from ggplot import *\n",
    "#heatingorsystemtypeid\n",
    "#heating_system_id\n",
    "\n",
    "\n",
    "#ggplot(aes(x='latitude', y='longitude', color='heatingorsystemtypeid'), data=prop2016) +  geom_point() +   scale_color_gradient(low = 'red', high = 'blue')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [],
   "source": [
    "#from ggplot import *\n",
    "#airconditioningtypeid\n",
    "#ac_id\n",
    "#ggplot(aes(x='latitude', y='longitude', color='airconditioningtypeid'), data=prop2016) +  geom_point() +   scale_color_gradient(low = 'red', high = 'blue')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "61494\n",
      "(90275, 45)\n",
      "61494\n",
      "(90273, 45)\n",
      "# of miss 61494\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   32.6s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "           metric_params=None, n_jobs=-1, n_neighbors=25, p=2,\n",
      "           weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 25, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(f1_score, pos_label=None, average=weighted)\n",
      "predicted output shape (61494,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "\n",
    "print(df_train.ac_id.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "temp['ac_id']=temp['ac_id'].fillna(-1)\n",
    "temp=temp.groupby(\"ac_id\").filter(lambda x: x.ac_id.size > 3)\n",
    "temp['ac_id'] = temp['ac_id'].replace(-1,np.nan)\n",
    "print(temp.ac_id.isnull().sum())\n",
    "print(temp.shape)\n",
    "\n",
    "missing_values=fillna_knn(temp,\n",
    "                  base = [ 'latitude', 'longitude' ] ,\n",
    "                  target = 'ac_id')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['ac_id'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'ac_id' ]  = missing_values\n",
    "\n",
    "\n",
    "print(df_train.ac_id.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "#from ggplot import *\n",
    "\n",
    "#ggplot(aes(x='latitude', y='longitude', color='yearbuilt'), data=df_train) +  geom_point() +   scale_color_gradient(low='#E1FA72', high='#F46FEE')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "756\n",
      "(90275, 45)\n",
      "756\n",
      "(90261, 45)\n",
      "# of miss 756\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:  1.2min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "           metric_params=None, n_jobs=-1, n_neighbors=5, p=2,\n",
      "           weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 5, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(f1_score, pos_label=None, average=weighted)\n",
      "predicted output shape (756,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#yearbuilt\n",
    "\n",
    "print(df_train.yearbuilt.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "temp['yearbuilt']=temp['yearbuilt'].fillna(-1)\n",
    "temp=temp.groupby(\"yearbuilt\").filter(lambda x: x.yearbuilt.size > 3)\n",
    "temp['yearbuilt'] = temp['yearbuilt'].replace(-1,np.nan)\n",
    "print(temp.yearbuilt.isnull().sum())\n",
    "print(temp.shape)\n",
    "\n",
    "missing_values=fillna_knn(temp,\n",
    "                  base = [ 'latitude', 'longitude','buildingqualitytypeid','propertylandusetypeid' ] ,\n",
    "                  target = 'yearbuilt')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['yearbuilt'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'yearbuilt' ]  = missing_values\n",
    "print(df_train.yearbuilt.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "#location seems to be related to building quality, (knnregressor)\n",
    "from sklearn.model_selection import KFold\n",
    "\n",
    "def fillna_knnr( df, base, target):\n",
    "    data_colnames = [ target ] + base\n",
    "    #print(\"data_colnames\",data_colnames)\n",
    "    missing_values_boolflag = df[target].isnull() #true for missing rows, false for columns with values\n",
    "    #print(\"miss\",missing_values_boolflag.head())\n",
    "    not_missing_boolflag = ~missing_values_boolflag \n",
    "    #print(\"not miss\",not_missing_boolflag.head())\n",
    "    number_of_missing_val = missing_values_boolflag.sum()\n",
    "    print(\"# of miss\",number_of_missing_val)\n",
    "    not_missing_rows = df.loc[ not_missing_boolflag, data_colnames]\n",
    "    #print(not_missing_rows.head())\n",
    "    Y = not_missing_rows[target]\n",
    "    X = not_missing_rows[base]\n",
    "    X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.20, random_state=3192)\n",
    "    metrics       = ['euclidean'] \n",
    "    weights       = ['distance'] \n",
    "    numNeighbors  = [5,10,15,20,25]\n",
    "    param_grid    = dict(metric=metrics,weights=weights,n_neighbors=numNeighbors)\n",
    "    cv            = KFold(n_splits=3,random_state=3192,shuffle=False) \n",
    "    grid = GridSearchCV(neighbors.KNeighborsRegressor(n_jobs=-1),param_grid=param_grid,cv=cv,scoring='neg_mean_absolute_error',refit=True,return_train_score=True,verbose=1,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "    grid.fit(X_train ,Y_train)\n",
    "    #print(\"grid.cv_results_\",grid.cv_results_)\n",
    "    print(\"grid.best_estimator_\",grid.best_estimator_)\n",
    "    print(\"grid.best_params_\",grid.best_params_)\n",
    "    print(\"grid.scorer_\",grid.scorer_)\n",
    "    #print(\"grid.n_splits_\",grid.n_splits_)\n",
    "    y_true, y_pred = Y_test, grid.predict(X_test) \n",
    "    Z = grid.predict(df.loc[missing_values_boolflag, base])\n",
    "    #df.loc[ missing_values_boolflag, target ]  = Z\n",
    "    return Z"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8920\n",
      "(90275, 45)\n",
      "8920\n",
      "(29647, 45)\n",
      "# of miss 8920\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   31.2s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "          metric_params=None, n_jobs=-1, n_neighbors=5, p=2,\n",
      "          weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 5, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(mean_absolute_error, greater_is_better=False)\n",
      "predicted output shape (8920,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#garage_sqft\n",
    "print(df_train.garage_sqft.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.loc[df_train.garagecarcnt>0,df_train.columns].copy()\n",
    "\n",
    "print(temp.garage_sqft.isnull().sum())\n",
    "print(temp.shape)\n",
    "\n",
    "missing_values=fillna_knnr(temp,\n",
    "                  base = [ 'latitude', 'longitude','garagecarcnt'] ,\n",
    "                  target = 'garage_sqft')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['garage_sqft'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'garage_sqft' ] = missing_values\n",
    "print(df_train.garage_sqft.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = df_train.drop('parcelid', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "#All the other columns with missing values seems to be  integer, will need regression to be imputed,\n",
    "#time to get categorical variables hot encoded\n",
    "\n",
    "#Identify numerical columns to produce a heatmap\n",
    "catcols = ['ac_id','buildingqualitytypeid','deck_flag','fips', 'heating_system_id','has_hottub_or_spa',\n",
    "          'just_hottub_or_spa', 'pool_with_spa_tub_yes','pool_with_spa_tub_no','propertylandusetypeid','basement_flag'\n",
    "          ,'fireplaceflag','taxdelinquencyflag']\n",
    "numcols = [x for x in df_train.columns if x not in catcols]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [],
   "source": [
    "#from ggplot import *\n",
    "\n",
    "#ggplot(aes(x='latitude', y='longitude', color='total_finished_living_area_sqft'), data=df_train) +  geom_point() +   scale_color_gradient(low='#E1FA72', high='#F46FEE')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "661\n",
      "(90275, 44)\n",
      "661\n",
      "(90275, 44)\n",
      "# of miss 661\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:  1.6min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "          metric_params=None, n_jobs=-1, n_neighbors=15, p=2,\n",
      "          weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 15, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(mean_absolute_error, greater_is_better=False)\n",
      "predicted output shape (661,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#total_finished_living_area_sqft\n",
    "\n",
    "print(df_train.total_finished_living_area_sqft.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "print(temp.total_finished_living_area_sqft.isnull().sum())\n",
    "print(temp.shape)\n",
    "missing_values=fillna_knnr(temp,\n",
    "                  base = [ 'latitude', 'longitude','basementsqft','numberofstories','poolcnt','garagecarcnt','garage_sqft','propertylandusetypeid'] ,\n",
    "                  target = 'total_finished_living_area_sqft')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['total_finished_living_area_sqft'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'total_finished_living_area_sqft' ] = missing_values\n",
    "print(df_train.total_finished_living_area_sqft.isnull().sum())\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1165\n",
      "(90275, 44)\n",
      "1165\n",
      "(90275, 44)\n",
      "# of miss 1165\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   37.5s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "          metric_params=None, n_jobs=-1, n_neighbors=25, p=2,\n",
      "          weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 25, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(mean_absolute_error, greater_is_better=False)\n",
      "predicted output shape (1165,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#total_bath\t1165\n",
    "#full_bath\t1182\n",
    "#half_bath\t1182\n",
    "#roomcnt\t1416\n",
    "#bedroomcnt\t1421\n",
    "\n",
    "#total_finished_living_area_sqft\n",
    "\n",
    "print(df_train.total_bath.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "print(temp.total_bath.isnull().sum())\n",
    "print(temp.shape)\n",
    "missing_values=fillna_knnr(temp,\n",
    "                  base = ['propertylandusetypeid','total_finished_living_area_sqft' ] ,\n",
    "                  target = 'total_bath')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['total_bath'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'total_bath' ] = missing_values\n",
    "print(df_train.total_bath.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [],
   "source": [
    "#drop half_bath and full bath, as there are only redundant values of total_bath\n",
    "df_train = df_train.drop(['full_bath','half_bath'], axis=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1421\n",
      "(90275, 42)\n",
      "1421\n",
      "(90275, 42)\n",
      "# of miss 1421\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   35.9s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "          metric_params=None, n_jobs=-1, n_neighbors=25, p=2,\n",
      "          weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 25, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(mean_absolute_error, greater_is_better=False)\n",
      "predicted output shape (1421,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#bedroomcnt\t1421\n",
    "\n",
    "print(df_train.bedroomcnt.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "print(temp.bedroomcnt.isnull().sum())\n",
    "print(temp.shape)\n",
    "missing_values=fillna_knnr(temp,\n",
    "                  base = ['propertylandusetypeid','total_finished_living_area_sqft','total_bath' ] ,\n",
    "                  target = 'bedroomcnt')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['bedroomcnt'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'bedroomcnt' ] = missing_values\n",
    "print(df_train.bedroomcnt.isnull().sum())\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train['total_bath']=df_train.total_bath.round(1)\n",
    "df_train['bedroomcnt']=df_train.bedroomcnt.round(1)\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [],
   "source": [
    "#recalculate roomcnt\t1416 as we have used imputation for total_bath and bedroomcnt\n",
    "\n",
    "df_train.loc[(df_train.roomcnt.isnull()),['roomcnt']]=df_train.total_bath + df_train.bedroomcnt\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(90275, 42)\n",
      "(90274, 42)\n"
     ]
    }
   ],
   "source": [
    "print(df_train.shape)\n",
    "df_train =df_train.loc[(df_train.total_parcel_tax.notnull()) & (df_train.land_tax.notnull()),df_train.columns]\n",
    "\n",
    "print(df_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10150\n",
      "(90274, 42)\n",
      "10150\n",
      "(90274, 42)\n",
      "# of miss 10150\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   38.6s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "          metric_params=None, n_jobs=-1, n_neighbors=15, p=2,\n",
      "          weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 15, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(mean_absolute_error, greater_is_better=False)\n",
      "predicted output shape (10150,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#lot_area_sqft\n",
    "print(df_train.lot_area_sqft.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "print(temp.lot_area_sqft.isnull().sum())\n",
    "print(temp.shape)\n",
    "missing_values=fillna_knnr(temp,\n",
    "                  base = ['latitude','longitude','propertylandusetypeid','total_finished_living_area_sqft','roomcnt','numberofstories' ] ,\n",
    "                  target = 'lot_area_sqft')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['lot_area_sqft'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'lot_area_sqft' ] = missing_values.round(2)\n",
    "print(df_train.lot_area_sqft.isnull().sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "379\n",
      "(90274, 42)\n",
      "379\n",
      "(90274, 42)\n",
      "# of miss 379\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   35.8s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "          metric_params=None, n_jobs=-1, n_neighbors=25, p=2,\n",
      "          weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 25, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(mean_absolute_error, greater_is_better=False)\n",
      "predicted output shape (379,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "# predict structure_tax and recalculate  total_parcel_tax = land_tax + structure_tax\n",
    "\n",
    "\n",
    "print(df_train.structure_tax.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "print(temp.structure_tax.isnull().sum())\n",
    "print(temp.shape)\n",
    "missing_values=fillna_knnr(temp,\n",
    "                  base = ['latitude','longitude','lot_area_sqft','propertylandusetypeid','total_finished_living_area_sqft','roomcnt','numberofstories' ] ,\n",
    "                  target = 'structure_tax')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['structure_tax'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'structure_tax' ] = missing_values.round(2)\n",
    "print(df_train.structure_tax.isnull().sum())\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [],
   "source": [
    "#36 total_property_tax_2016 \n",
    "\n",
    "#total_parcel_tax = land_tax + structure_tax\n",
    "    \n",
    "df_train['total_parcel_tax']=df_train['structure_tax']+df_train['land_tax']\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [],
   "source": [
    "#age of the property\n",
    "df_train['age'] = 2016 - df_train['yearbuilt']\n",
    "df_train=df_train.drop(['yearbuilt'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n",
      "(90274, 42)\n",
      "6\n",
      "(90274, 42)\n",
      "# of miss 6\n",
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:   37.4s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "grid.best_estimator_ KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='euclidean',\n",
      "          metric_params=None, n_jobs=-1, n_neighbors=25, p=2,\n",
      "          weights='distance')\n",
      "grid.best_params_ {'metric': 'euclidean', 'n_neighbors': 25, 'weights': 'distance'}\n",
      "grid.scorer_ make_scorer(mean_absolute_error, greater_is_better=False)\n",
      "predicted output shape (6,)\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "#total_property_tax_2016\n",
    "\n",
    "\n",
    "print(df_train.total_property_tax_2016.isnull().sum())\n",
    "print(df_train.shape)\n",
    "temp=df_train.copy()\n",
    "print(temp.total_property_tax_2016.isnull().sum())\n",
    "print(temp.shape)\n",
    "missing_values=fillna_knnr(temp,\n",
    "                  base = ['latitude','longitude','lot_area_sqft','propertylandusetypeid','total_finished_living_area_sqft','roomcnt','numberofstories' ] ,\n",
    "                  target = 'total_property_tax_2016')\n",
    "\n",
    "print(\"predicted output shape\",missing_values.shape)\n",
    "missing_values_boolflag = df_train['total_property_tax_2016'].isnull()\n",
    "df_train.loc[ missing_values_boolflag, 'total_property_tax_2016' ] = missing_values.round(2)\n",
    "print(df_train.total_property_tax_2016.isnull().sum())\n",
    "    \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                             column_name  missing_count\n",
      "7  finished_living_area_entryfloor_sqft2          83419\n",
      "9  finished_living_area_entryfloor_sqft1          83419\n",
      "(2, 2)\n"
     ]
    }
   ],
   "source": [
    "#check missing values\n",
    "\n",
    "missing_df = df_train.isnull().sum(axis=0).reset_index()\n",
    "missing_df.columns = ['column_name', 'missing_count']\n",
    "missing_df = missing_df.loc[missing_df['missing_count']>0]\n",
    "missing_df = missing_df.sort_values(by='missing_count')\n",
    "print(missing_df)\n",
    "print(missing_df.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [],
   "source": [
    "#both the columns above miss 92% of the data, there is no related varibale to impute it, hence dropping them at this point\n",
    "\n",
    "df_train = df_train.drop(['finished_living_area_entryfloor_sqft2','finished_living_area_entryfloor_sqft1'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Identify numerical columns to produce a heatmap\n",
    "catcols = ['ac_id','buildingqualitytypeid','deck_flag','fips','pool_with_spa_tub_no','pool_with_spa_tub_yes','has_hottub_or_spa',\n",
    "           'just_hottub_or_spa','heating_system_id','propertylandusetypeid','basement_flag','fireplaceflag','taxdelinquencyflag']\n",
    "numcols = [x for x in df_train.columns if x not in catcols]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 2 variables are in object datatype, coverting into numeric\n",
    "df_train[['census_tractnumber','block_number']] = df_train[['census_tractnumber','block_number']].apply(pd.to_numeric)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dropping categorical columns as xgboost feature selection cannot hadle it\n",
    "\n",
    "train_x = df_train.drop(catcols+['logerror'], axis=1)\n",
    "\n",
    "train_y=df_train['logerror']\n",
    "\n",
    "train_x = train_x.astype(float) \n",
    "train_y = train_y.astype(float)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.options.display.max_rows = 65\n",
    "\n",
    "dtype_df = train_x.dtypes.reset_index()\n",
    "dtype_df.columns = [\"Count\", \"Column Type\"]\n",
    "#dtype_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train.loc[df_train.has_hottub_or_spa==True,'has_hottub_or_spa']=\"Yes\"\n",
    "df_train.loc[df_train.has_hottub_or_spa==0,'has_hottub_or_spa']=\"No\"\n",
    "\n",
    "df_train.loc[df_train.just_hottub_or_spa==0,'just_hottub_or_spa']=\"No\"\n",
    "df_train.loc[df_train.just_hottub_or_spa==1,'just_hottub_or_spa']=\"Yes\"\n",
    "\n",
    "df_train.loc[df_train.deck_flag==0,'deck_flag']=\"No\"\n",
    "df_train.loc[df_train.deck_flag==1,'deck_flag']=\"Yes\"\n",
    "\n",
    "df_train.loc[df_train.basement_flag==0,'basement_flag']=\"No\"\n",
    "df_train.loc[df_train.basement_flag==1,'basement_flag']=\"Yes\"\n",
    "\n",
    "df_train.loc[df_train.fireplaceflag==False,'fireplaceflag']=\"No\"\n",
    "df_train.loc[df_train.fireplaceflag==True,'fireplaceflag']=\"Yes\"\n",
    "#"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [],
   "source": [
    "#ac_id,heating_system_id,propertylandusetypeid\n",
    "dummieslist=['has_hottub_or_spa','just_hottub_or_spa',\n",
    "             'deck_flag','fips','basement_flag','fireplaceflag','taxdelinquencyflag']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(90274, 15)"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train[dummieslist] = df_train[dummieslist].astype(object)\n",
    "dummies = pd.get_dummies(df_train[dummieslist], prefix= dummieslist)\n",
    "dummies.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [],
   "source": [
    "dummies2=['pool_with_spa_tub_no','pool_with_spa_tub_yes']\n",
    "df_train[dummies2] = df_train[dummies2].astype(int)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sqlalchemy\\dialects\\mysql\\base.py:1569: Warning: (1287, \"'@@tx_isolation' is deprecated and will be removed in a future release. Please use '@@transaction_isolation' instead\")\n",
      "  cursor.execute('SELECT @@tx_isolation')\n"
     ]
    },
    {
     "ename": "NameError",
     "evalue": "name 'df_train' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-1-d32a1095d7fe>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      6\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      7\u001b[0m \u001b[1;32mwith\u001b[0m \u001b[0mengine\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mconnect\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0mcon\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcon\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbegin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 8\u001b[1;33m     \u001b[0mdf_train\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mto_sql\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'df_train_f1'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mengine\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mchunksize\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m10000\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mindex\u001b[0m \u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mif_exists\u001b[0m \u001b[1;33m=\u001b[0m\u001b[1;34m'replace'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'df_train' is not defined"
     ]
    }
   ],
   "source": [
    "import MySQLdb\n",
    "from sqlalchemy import create_engine\n",
    "engineString = 'mysql+mysqldb://root:MyNewPass@localhost/sakila'\n",
    "engine = create_engine(engineString)\n",
    "con=engine.connect()\n",
    "\n",
    "with engine.connect() as con, con.begin():\n",
    "    df_train.to_sql('df_train_f1', engine, chunksize=10000, index =False,if_exists ='replace')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sqlalchemy\\dialects\\mysql\\base.py:1569: Warning: (1287, \"'@@tx_isolation' is deprecated and will be removed in a future release. Please use '@@transaction_isolation' instead\")\n",
      "  cursor.execute('SELECT @@tx_isolation')\n"
     ]
    }
   ],
   "source": [
    "import MySQLdb\n",
    "from sqlalchemy import create_engine\n",
    "engineString = 'mysql+mysqldb://root:MyNewPass@localhost/sakila'\n",
    "engine = create_engine(engineString)\n",
    "con=engine.connect()\n",
    "df_train=pd.read_sql_table('df_train_f1',con=engine)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "numcols2=['basementsqft','total_bath','bedroomcnt','total_finished_living_area_sqft','fireplace_count','garagecarcnt',\n",
    " 'garage_sqft','latitude','longitude','lot_area_sqft','poolcnt','pool_sqft','roomcnt','unitcnt','patio_sqft','storage_sqft',\n",
    " 'numberofstories','structure_tax','total_parcel_tax','land_tax','total_property_tax_2016','taxdelinquencyyear','transaction_month',\n",
    " 'census_tractnumber','block_number','age']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y=df_train['logerror']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#buildingqualitytypeid ->has order\n",
    "le = LabelEncoder()\n",
    "df_train['buildingqualitytypeid']=le.fit_transform(df_train.buildingqualitytypeid)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df_train.ac_id.value_counts()\n",
    "#df_train.propertylandusetypeid.value_counts()\n",
    "#'buildingqualitytypeid','ac_id','heating_system_id','propertylandusetypeid'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'dummies' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-9-b0b091d234fc>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mX\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mpd\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mconcat\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mdummies\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mdf_train\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mdummies2\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mdf_train\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mnumcols2\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mdf_train\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'buildingqualitytypeid'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'ac_id'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'heating_system_id'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'propertylandusetypeid'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      2\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'dummies' is not defined"
     ]
    }
   ],
   "source": [
    "X=pd.concat([dummies,df_train[dummies2],df_train[numcols2],df_train[['buildingqualitytypeid','ac_id','heating_system_id','propertylandusetypeid']]],axis=1)\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'X' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-10-7e8785b34441>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m \u001b[0mX_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX_test\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mY_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mY_test\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtrain_test_split\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mY\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtest_size\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0.10\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mrandom_state\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m3192\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'X' is not defined"
     ]
    }
   ],
   "source": [
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.10, random_state=3192)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0kAAAQMCAYAAABDS9IRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAACH/wNzy2HzzvXYjgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x24b01ba3390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#  top features\n",
    "import xgboost as xgb\n",
    "xgb_params = {\n",
    "    'eta': 0.05,\n",
    "    'max_depth': 8,\n",
    "    'subsample': 0.7,\n",
    "    'colsample_bytree': 0.7,\n",
    "    'objective': 'reg:linear',\n",
    "    'silent': 1,\n",
    "    'seed' : 0\n",
    "}\n",
    "dtrain = xgb.DMatrix(X_train, Y_train, feature_names=X_train.columns.values)\n",
    "model = xgb.train(dict(xgb_params, silent=0), dtrain, num_boost_round=50)\n",
    "# plot the important features #\n",
    "fig, ax = plt.subplots(figsize=(12,18))\n",
    "#max_num_features=50, error for no reason \n",
    "xgb.plot_importance(model, height=0.8, ax=ax)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [],
   "source": [
    "#top features\n",
    "xgboost_selection=['total_finished_living_area_sqft','latitude','structure_tax','total_property_tax_2016',\n",
    "'total_parcel_tax','land_tax','longitude','lot_area_sqft','census_tractnumber','age','total_bath','bedroomcnt',\n",
    "'block_number','transaction_month','roomcnt','taxdelinquencyyear','unitcnt','taxdelinquencyflag_No',\n",
    "'fips_LA','garage_sqft','pool_with_spa_tub_no','has_hottub_or_spa_No','garagecarcnt','deck_flag_No',\n",
    "'poolcnt','pool_sqft'\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(81246, 47)\n",
      "(81246, 21)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtEAAAM3CAYAAAD7owwsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x24b27258b70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# feature selection\n",
    "#c_id,heating_system_id,propertylandusetypeid\n",
    "from sklearn.ensemble import ExtraTreesRegressor\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "reg = ExtraTreesRegressor(n_estimators=500, max_depth=8, max_features='sqrt',\n",
    "                          min_samples_split=100 ,min_samples_leaf=10, bootstrap=True,n_jobs=-1, random_state=3192)\n",
    "reg = reg.fit(X_train, Y_train)\n",
    "#print(\"importance\",reg.feature_importances_) \n",
    "model = SelectFromModel(reg, prefit=True)\n",
    "X_new = model.transform(X_train)\n",
    "print(X_train.shape)\n",
    "print(X_new.shape)  \n",
    "\n",
    "feat_names = X.columns.values\n",
    "importances = reg.feature_importances_\n",
    "std = np.std([tree.feature_importances_ for tree in reg.estimators_], axis=0)\n",
    "indices = np.argsort(importances)[::-1][:26]\n",
    "plt.figure(figsize=(12,12))\n",
    "plt.title(\"Feature importances\")\n",
    "plt.bar(range(len(indices)), importances[indices], color=\"r\", yerr=std[indices], align=\"center\")\n",
    "plt.xticks(range(len(indices)), feat_names[indices], rotation='vertical')\n",
    "plt.xlim([-1, len(indices)])\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [],
   "source": [
    "tree_selection=[\n",
    "    'total_finished_living_area_sqft','structure_tax','total_property_tax_2016','total_bath','total_parcel_tax',\n",
    "    'age','latitude','census_tractnumber','bedroomcnt','longitude','land_tax','propertylandusetypeid','block_number',\n",
    "    'buildingqualitytypeid','numberofstories','heating_system_id','unitcnt','transaction_month','lot_area_sqft','roomcnt',\n",
    "    'garage_sqft','garagecarcnt','pool_with_spa_tub_no','poolcnt','fips_LA','taxdelinquencyyear','patio_sqft',\n",
    "    'taxdelinquencyflag_No','taxdelinquencyflag_Yes'\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [],
   "source": [
    "#print(\"Feature ranking:\")\n",
    "\n",
    "#for f in range(X.shape[1]):\n",
    "#    print(\"%d. feature %d (%f)\" % (f + 1, indices[f], importances[indices[f]]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of features : 43\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaUAAAEWCAYAAADGjIh1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSoIgCMJTQyQlQRAE4anx/yA5BsoKqchcAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x24b262cdc50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.linear_model import Ridge,Lasso\n",
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer\n",
    "\n",
    "#model=Lasso(alpha=0.2, fit_intercept=True, normalize=True, precompute=False, copy_X=True,\n",
    " #                                max_iter=1000, \n",
    "  #                               tol=0.0001, warm_start=False, positive=False, random_state=3192, selection='cyclic')\n",
    "\n",
    "#Ridge(random_state=3192,solver='auto',fit_intercept=True,normalize=True,alpha=0.1)\n",
    "#LinearRegression(n_jobs=-1,fit_intercept=True, normalize=True, copy_X=True)\n",
    "\n",
    "\n",
    "rfecv = RFECV(estimator=LinearRegression(n_jobs=-1,fit_intercept=True, normalize=True, copy_X=True), step=2, cv=KFold(4),scoring='neg_mean_absolute_error')\n",
    "rfecv.fit(X_train, Y_train)\n",
    "\n",
    "print(\"Optimal number of features : %d\" % rfecv.n_features_)\n",
    "\n",
    "# Plot number of features VS. cross-validation scores\n",
    "plt.figure()\n",
    "plt.xlabel(\"Number of features selected\")\n",
    "\n",
    "plt.ylabel(\"Cross validation score (nb of correct classifications)\")\n",
    "plt.plot(range(1, len(rfecv.grid_scores_) + 1), rfecv.grid_scores_)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfe_selection = [i for indx,i in enumerate(X.columns) if rfecv.support_[indx] == True]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE on test data 0.0674343856764\n",
      "MSE on test data 0.0235487116844\n",
      "R Squared data  0.00158156841316\n"
     ]
    }
   ],
   "source": [
    "#Linear regression with rfe_selection selection\n",
    "#rfe_selection, tree_selection, xgboost_selection\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer,mean_squared_error\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "parameters = {'fit_intercept':[True], 'normalize':[True,False], 'copy_X':[True]}\n",
    "scoring = {'MAE':'neg_mean_absolute_error','MSE': make_scorer(mean_squared_error,greater_is_better=False)}\n",
    "\n",
    "grid1 = GridSearchCV(LinearRegression(n_jobs=-1),param_grid=parameters, scoring=scoring,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid1.fit(X_train[rfe_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "Y_pred = grid1.predict(X_test[rfe_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared data \",r2_score(Y_test,Y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE on test data 0.0672656713421\n",
      "MSE on test data 0.0234451555385\n",
      "R Squared data  0.00597214256383\n"
     ]
    }
   ],
   "source": [
    "#tree_selection\n",
    "\n",
    "#Linear regression with tree_selection selection\n",
    "#rfe_selection, tree_selection, xgboost_selection\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "parameters = {'fit_intercept':[True], 'normalize':[True,False], 'copy_X':[True]}\n",
    "scoring = {'MAE':'neg_mean_absolute_error','MSE': make_scorer(mean_squared_error,greater_is_better=False)}\n",
    "\n",
    "grid2 = GridSearchCV(LinearRegression(n_jobs=-1),param_grid=parameters, scoring=scoring,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid2.fit(X_train[tree_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "Y_pred = grid2.predict(X_test[tree_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared data \",r2_score(Y_test,Y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE on test data 0.0672783506936\n",
      "MSE on test data 0.0234485338833\n",
      "R Squared data  0.0058289074789\n"
     ]
    }
   ],
   "source": [
    " #xgboost_selection\n",
    "\n",
    "#Linear regression with tree_selection selection\n",
    "#rfe_selection, tree_selection, xgboost_selection\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "parameters = {'fit_intercept':[True], 'normalize':[True,False], 'copy_X':[True]}\n",
    "scoring = {'MAE':'neg_mean_absolute_error','MSE': make_scorer(mean_squared_error,greater_is_better=False)}\n",
    "\n",
    "grid3 = GridSearchCV(LinearRegression(n_jobs=-1),param_grid=parameters, scoring=scoring,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid3.fit(X_train[xgboost_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "Y_pred = grid3.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared data \",r2_score(Y_test,Y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best params {'alpha': 0.2, 'copy_X': True, 'fit_intercept': True, 'normalize': True}\n",
      "MAE on test data 0.0673806454705\n",
      "MSE on test data 0.0235434737612\n",
      "R Squared data  0.0018036459175\n"
     ]
    }
   ],
   "source": [
    "#Ridge\n",
    "#rfe_selection, tree_selection, xgboost_selection\n",
    "from sklearn.linear_model import Ridge\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from scipy.stats import uniform as sp_rand\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "parameters = {'fit_intercept':[True], 'normalize':[True], 'copy_X':[True],'alpha':[0.001,0.1,0.01,0.02,0.03,0.05,0.2,0.015,0.020]}\n",
    "scoring = {'MAE':'neg_mean_absolute_error','MSE': make_scorer(mean_squared_error,greater_is_better=False)}\n",
    "\n",
    "grid4 = GridSearchCV(Ridge(random_state=3192,solver='auto'),param_grid=parameters, scoring=scoring,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid4.fit(X_train[rfe_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "print(\"best params\",grid4.best_params_)\n",
    "Y_pred = grid4.predict(X_test[rfe_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared data \",r2_score(Y_test,Y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best params {'alpha': 0.1, 'copy_X': True, 'fit_intercept': True, 'normalize': True}\n",
      "MAE on test data 0.0672249893864\n",
      "MSE on test data 0.0234815834045\n",
      "R Squared data  0.00442767366169\n"
     ]
    }
   ],
   "source": [
    "#Ridge\n",
    "#rfe_selection, tree_selection, xgboost_selection\n",
    "from sklearn.linear_model import Ridge\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from scipy.stats import uniform as sp_rand\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "parameters = {'fit_intercept':[True], 'normalize':[True], 'copy_X':[True],'alpha':[0.001,0.1,0.01,0.02,0.03,0.05,0.2,0.015,0.020]}\n",
    "scoring = {'MAE':'neg_mean_absolute_error','MSE': make_scorer(mean_squared_error,greater_is_better=False)}\n",
    "\n",
    "grid5 = GridSearchCV(Ridge(random_state=3192,solver='auto'),param_grid=parameters, scoring=scoring,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid5.fit(X_train[tree_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "print(\"best params\",grid5.best_params_)\n",
    "Y_pred = grid5.predict(X_test[tree_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared data \",r2_score(Y_test,Y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best params {'alpha': 0.1, 'copy_X': True, 'fit_intercept': True, 'normalize': True}\n",
      "MAE on test data 0.0672222537256\n",
      "MSE on test data 0.0234780234248\n",
      "R Squared data  0.00457860970566\n"
     ]
    }
   ],
   "source": [
    "#Ridge\n",
    "#rfe_selection, tree_selection, xgboost_selection\n",
    "from sklearn.linear_model import Ridge\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from scipy.stats import uniform as sp_rand\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "parameters = {'fit_intercept':[True], 'normalize':[True], 'copy_X':[True],'alpha':[0.001,0.1,0.01,0.02,0.03,0.05,0.2,0.015]}\n",
    "scoring = {'MAE':'neg_mean_absolute_error','MSE': make_scorer(mean_squared_error,greater_is_better=False)}\n",
    "\n",
    "grid6 = GridSearchCV(Ridge(random_state=3192,solver='auto'),param_grid=parameters, scoring=scoring,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid6.fit(X_train[xgboost_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "print(\"best params\",grid6.best_params_)\n",
    "Y_pred = grid6.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared data \",r2_score(Y_test,Y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best params {'alpha': 0.001, 'copy_X': True, 'fit_intercept': True, 'l1_ratio': 0, 'normalize': True}\n",
      "MAE on test data 0.0675360249797\n",
      "MSE on test data 0.0235838967882\n",
      "R Squared 8.97901549788e-05\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\coordinate_descent.py:491: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations. Fitting data with very small alpha may cause precision problems.\n",
      "  ConvergenceWarning)\n"
     ]
    }
   ],
   "source": [
    "#ElasticNet\n",
    "#rfe_selection, tree_selection, xgboost_selection\n",
    "from sklearn.linear_model import ElasticNet\n",
    "from scipy.stats import uniform as sp_rand\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "parameters = {'l1_ratio':[0],'fit_intercept':[True], 'normalize':[True], 'copy_X':[True],'alpha':[0.001,0.1,0.01,0.02,0.03,0.05,0.2,0.015]}\n",
    "scoring = {'MAE':'neg_mean_absolute_error','MSE': make_scorer(mean_squared_error,greater_is_better=False)}\n",
    "\n",
    "grid7 = GridSearchCV(ElasticNet(positive=False,random_state=3192),param_grid=parameters, scoring=scoring,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid7.fit(X_train[xgboost_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "print(\"best params\",grid7.best_params_)\n",
    "Y_pred = grid7.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared\",r2_score(Y_test,Y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best params {'n_estimators': 80}\n",
      "MAE on test data 0.0673119750138\n",
      "MSE on test data 0.0231574575868\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.metrics import mean_absolute_error,make_scorer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "\n",
    "\n",
    "parameters_n_estimators = {'n_estimators':[80,100]}\n",
    "              \n",
    "              \n",
    "\n",
    "              \n",
    "grid8 = GridSearchCV(GradientBoostingRegressor(loss='ls',criterion='friedman_mse',learning_rate=0.1,\n",
    "                                             subsample=0.8,min_samples_split=2000,min_samples_leaf=50,\n",
    "                                             max_depth=8,max_features='sqrt',random_state=3192),param_grid=parameters_n_estimators, \n",
    "                    cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid8.fit(X_train[xgboost_selection], Y_train)\n",
    "#print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "print(\"best params\",grid8.best_params_)\n",
    "Y_pred = grid8.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5. grid best_score_ 0.0132867697025\n",
      "best params {'n_estimators': 160}\n",
      "best score 0.0132867697025\n",
      "MAE on test data 0.0674965531466\n",
      "MSE on test data 0.023164925202\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.metrics import mean_absolute_error,make_scorer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "\n",
    "param_grid={'n_estimators':[160,200]}\n",
    "              \n",
    "grid9 = GridSearchCV(GradientBoostingRegressor(subsample=0.8,max_features=11,min_samples_leaf=50,min_samples_split=50,\n",
    "                                               \n",
    "                                               max_depth=9,loss='ls',criterion='friedman_mse',learning_rate=0.05,\n",
    "                                             random_state=3192), param_grid=param_grid,cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid9.fit(X_train[xgboost_selection], Y_train)\n",
    "print(\"5. grid best_score_\",abs(grid9.best_score_))\n",
    "print(\"best params\",grid9.best_params_)\n",
    "print(\"best score\",grid9.best_score_)\n",
    "Y_pred = grid9.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5. grid best_score_ 0.0154254980517\n",
      "best params {'n_estimators': 600}\n",
      "best score 0.0154254980517\n",
      "MAE on test data 0.0672418061984\n",
      "MSE on test data 0.0231165901657\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.metrics import mean_absolute_error,make_scorer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "\n",
    " #0.01 for 600 trees\n",
    "param_grid={'n_estimators':[400,600]}\n",
    "\n",
    "              \n",
    "grid10 = GridSearchCV(GradientBoostingRegressor(subsample=0.8,max_features=11,min_samples_leaf=50,min_samples_split=50,\n",
    "                                               \n",
    "                                               max_depth=9,loss='ls',criterion='friedman_mse',learning_rate=0.01,\n",
    "                                             random_state=3192),\n",
    "                     param_grid=param_grid, cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid10.fit(X_train[xgboost_selection], Y_train)\n",
    "print(\"5. grid best_score_\",abs(grid10.best_score_))\n",
    "print(\"best params\",grid10.best_params_)\n",
    "print(\"best score\",grid10.best_score_)\n",
    "Y_pred = grid10.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5. grid best_score_ 0.015480437058\n",
      "best params {'n_estimators': 1200}\n",
      "best score 0.015480437058\n",
      "MAE on test data 0.0671656756254\n",
      "MSE on test data 0.0231048416185\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.metrics import mean_absolute_error,make_scorer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "\n",
    " # 0.005 for 1200 trees.\n",
    "#param_grid={'n_estimators':[1000,1200]}\n",
    "param_grid={'n_estimators':[1200]}              \n",
    "grid11 = GridSearchCV(GradientBoostingRegressor(subsample=0.8,max_features=11,min_samples_leaf=50,min_samples_split=50,\n",
    "                                        \n",
    "                                               max_depth=9,loss='ls',criterion='friedman_mse',learning_rate=0.005,\n",
    "                                             random_state=3192),\n",
    "                     param_grid=param_grid, cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid11.fit(X_train[xgboost_selection], Y_train)\n",
    "print(\"5. grid best_score_\",abs(grid11.best_score_))\n",
    "print(\"best params\",grid11.best_params_)\n",
    "print(\"best score\",grid11.best_score_)\n",
    "Y_pred = grid11.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5. grid best_score_ 0.0160962097719\n",
      "best params {'max_features': 22, 'n_estimators': 1200}\n",
      "best score 0.0160962097719\n",
      "MAE on test data 0.0671686272257\n",
      "MSE on test data 0.0230879788586\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.metrics import mean_absolute_error,make_scorer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "\n",
    " # 0.005 for 1200 trees.\n",
    "param_grid={'n_estimators':[1200],'max_features':[22]}\n",
    "\n",
    "              \n",
    "grid12 = GridSearchCV(GradientBoostingRegressor(subsample=0.8,min_samples_leaf=50,min_samples_split=50,max_depth=9,loss='ls',criterion='friedman_mse',learning_rate=0.005,random_state=3192),\n",
    "                     param_grid=param_grid, cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=2,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid12.fit(X_train, Y_train)\n",
    "print(\"5. grid best_score_\",abs(grid12.best_score_))\n",
    "print(\"best params\",grid12.best_params_)\n",
    "print(\"best score\",grid12.best_score_)\n",
    "Y_pred = grid12.predict(X_test)\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE 0.0670942557879\n"
     ]
    }
   ],
   "source": [
    "#pending\n",
    "#from  xgboost import XGBRegressor\n",
    "#xgboost.XGBRegressor\n",
    "model12 = XGBRegressor()\n",
    "model12.fit(X_train, Y_train)\n",
    "# make predictions for test data\n",
    "y_pred = model12.predict(X_test)\n",
    "print(\"MAE\",mean_absolute_error(Y_test, y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PCA(copy=True, iterated_power='auto', n_components=None, random_state=None,\n",
       "  svd_solver='auto', tol=0.0, whiten=False)"
      ]
     },
     "execution_count": 190,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#pca selection\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.preprocessing import scale\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.preprocessing import scale\n",
    "%matplotlib inline\n",
    "scaled_x = scale(X)\n",
    "pca = PCA(n_components=None, copy=True, whiten=False, svd_solver='auto', tol=0.0, iterated_power='auto', random_state=None)\n",
    "pca.fit(scaled_x)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[  15.45   26.55   35.96   42.71   48.82   54.77   60.56   65.01   68.66\n",
      "   72.22   75.07   77.35   79.51   81.64   83.72   85.77   87.56   89.06\n",
      "   90.45   91.75   93.03   94.21   95.16   95.95   96.63   97.2    97.71\n",
      "   98.2    98.61   98.94   99.24   99.52   99.72   99.87   99.96  100.\n",
      "  100.01  100.01  100.01  100.01  100.01  100.01  100.01  100.01  100.01\n",
      "  100.01  100.01]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x24b63b37630>]"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD3CAYAAAANMK+RAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEQp0ERGP+P85AeTERB+noQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x24b272636d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "#The amount of variance that each PC explains\n",
    "var= pca.explained_variance_ratio_\n",
    "#Cumulative Variance explains\n",
    "var1=np.cumsum(np.round(pca.explained_variance_ratio_, decimals=4)*100)\n",
    "print(var1)\n",
    "plt.plot(var1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Looking at above plot I'm taking 28 variables\n",
    "\n",
    "pca = PCA(n_components=28, copy=True, whiten=False, svd_solver='auto', tol=0.0, iterated_power='auto', random_state=None)\n",
    "pca.fit(scaled_x)\n",
    "\n",
    "pca1=pca.fit_transform(scaled_x)\n",
    "\n",
    "pca = PCA(n_components=28, copy=True, whiten=True, svd_solver='auto', tol=0.0, iterated_power='auto', random_state=None)\n",
    "pca.fit(scaled_x)\n",
    "pca2=pca.fit_transform(scaled_x)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "pcaX_train, pcaX_test, pcaY_train, pcaY_test = train_test_split(pca1, Y, test_size=0.10, random_state=3192)\n",
    "pca2X_train, pca2X_test, pca2Y_train, pca2Y_test = train_test_split(pca2, Y, test_size=0.10, random_state=3192)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done   5 out of   5 | elapsed: 53.7min remaining:    0.0s\n",
      "[Parallel(n_jobs=-1)]: Done   5 out of   5 | elapsed: 53.7min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5. grid best_score_ 0.00377417855648\n",
      "best params {'max_features': 22, 'n_estimators': 1200}\n",
      "best score 0.00377417855648\n",
      "MAE on test data 0.0677470539474\n",
      "MSE on test data 0.023465957039\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.metrics import mean_absolute_error,make_scorer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# just to check whether normalized /not normalized data gives better results\n",
    "\n",
    " # 0.005 for 1200 trees.\n",
    "param_grid={'n_estimators':[1200],'max_features':[22]}\n",
    "\n",
    "              \n",
    "grid13 = GridSearchCV(GradientBoostingRegressor(subsample=0.8,min_samples_leaf=50,min_samples_split=50,max_depth=9,loss='ls',criterion='friedman_mse',learning_rate=0.005,random_state=3192),\n",
    "                     param_grid=param_grid, cv=5,refit='MAE',\n",
    "                    return_train_score=True,\n",
    "                    verbose=2,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid13.fit(pcaX_train, pcaY_train)\n",
    "print(\"5. grid best_score_\",abs(grid13.best_score_))\n",
    "print(\"best params\",grid13.best_params_)\n",
    "print(\"best score\",grid13.best_score_)\n",
    "Y_pred = grid13.predict(pcaX_test)\n",
    "print(\"MAE on test data\",mean_absolute_error(pcaY_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(pcaY_test,Y_pred))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "def modelfit(model, X_train, Y_train,X_test,Y_test,useTrainCV=True, cv_folds=5, early_stopping_rounds=50):\n",
    "    \n",
    "    if useTrainCV:\n",
    "        xgb_param = model.get_xgb_params()\n",
    "        \n",
    "        xgtrain = xgb.DMatrix(X_train, Y_train, feature_names=X_train.columns.values)\n",
    "        \n",
    "        cvresult = xgb.cv(xgb_param, xgtrain, num_boost_round=model.get_params()['n_estimators'], nfold=cv_folds,\n",
    "            metrics='mae', early_stopping_rounds=early_stopping_rounds, verbose_eval=1,seed=3192)\n",
    "        \n",
    "        model.set_params(n_estimators=cvresult.shape[0])\n",
    "    \n",
    "    #Fit the algorithm on the data\n",
    "    model.fit(X_train,Y_train,eval_metric='mae')\n",
    "        \n",
    "    #Predict training set:\n",
    "    y_pred = model.predict(X_test)\n",
    "    print(\"MAE on test data\",mean_absolute_error(Y_test, y_pred))\n",
    "\n",
    "                    \n",
    "    feat_imp = pd.Series(model.booster().get_fscore()).sort_values(ascending=False)\n",
    "    feat_imp.plot(kind='bar', title='Feature Importances')\n",
    "    plt.ylabel('Feature Importance Score')\n",
    "    #print(\" model.get_params()['n_estimators']\",model.get_params()['n_estimators'])\n",
    "    print(\"n_estimators\",cvresult.shape[0])\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 11 candidates, totalling 55 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  33 tasks      | elapsed:  5.4min\n",
      "[Parallel(n_jobs=-1)]: Done  55 out of  55 | elapsed:  8.2min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5. grid best_score_ 0.0681882269452\n",
      "best params {'reg_alpha': 0.1}\n",
      "best score -0.0681882269452\n",
      "MAE on test data 0.0670225939664\n",
      "MSE on test data 0.0233410419075\n"
     ]
    }
   ],
   "source": [
    "param_test7 = {\n",
    " 'reg_alpha':[0.001,0.005,0.01,0.03,0.04,0.05,0.06,0.07,0.08,0.09,0.1]\n",
    "}\n",
    "\n",
    "gsearch7 = GridSearchCV(estimator = XGBRegressor(learning_rate =0.1,n_estimators=81,\n",
    "                    max_depth=2,\n",
    "                    min_child_weight=350,\n",
    "                    gamma=0.4,\n",
    "                    objective= 'reg:linear',\n",
    "                    nthread=-1,scale_pos_weight=1,seed=3192,colsample_bytree=0.9,subsample=0.95),\n",
    "param_grid = param_test7, scoring='neg_mean_absolute_error',n_jobs=-1, cv=5,verbose=2)\n",
    "\n",
    "gsearch7.fit(X_train, Y_train)\n",
    "print(\"5. grid best_score_\",abs(gsearch7.best_score_))\n",
    "print(\"best params\",gsearch7.best_params_)\n",
    "print(\"best score\",gsearch7.best_score_)\n",
    "\n",
    "Y_pred = gsearch7.predict(X_test)\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#0.05 0.0670037948915\n",
    "#0.0670416222198\n",
    "\n",
    "xgb24=XGBRegressor(learning_rate =0.1,n_estimators=1000,\n",
    "                    max_depth=2,\n",
    "                    min_child_weight=350,\n",
    "                    gamma=0.4,\n",
    "                    objective= 'reg:linear',\n",
    "                    nthread=-1,scale_pos_weight=1,seed=3192,colsample_bytree=0.9,subsample=0.95,reg_alpha=0.05)\n",
    "\n",
    "#recaliberate n_estimators with the values found above\n",
    "\n",
    "modelfit(xgb24, X_train, Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\grid_search.py:42: 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. This module will be removed in 0.20.\n",
      "  DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.datasets import make_regression\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.metrics import r2_score,mean_absolute_error,make_scorer,mean_squared_error\n",
    "\n",
    "\n",
    "regr = RandomForestRegressor(n_jobs=-1, random_state=30, verbose=2)\n",
    "parameters = { \"n_estimators\": [50],\"criterion\":[\"mae\"]}\n",
    "\n",
    "\n",
    "grid = GridSearchCV(regr,param_grid=parameters,cv=5,refit='MAE',verbose=0,n_jobs=-1,pre_dispatch='n_jobs')\n",
    "\n",
    "grid.fit(X_train[xgboost_selection], Y_train)\n",
    "print(\"5. grid best_score_\",abs(grid.best_score_))\n",
    "Y_pred = grid.predict(X_test[xgboost_selection])\n",
    "print(\"MAE on test data\",mean_absolute_error(Y_test,Y_pred))\n",
    "print(\"MSE on test data\",mean_squared_error(Y_test,Y_pred))\n",
    "print(\"R Squared data \",r2_score(Y_test,Y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 444,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW8AAAEFCAYAAAAsU2YoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x22318911278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Interpretation of Regression Plots #Take a look at the residual vs fitted values plot. #residual plot\n",
    "x_plot = plt.scatter(y_pred, (y_pred - Y_test), c='b')\n",
    "plt.hlines(y=0, xmin= -1000, xmax=5000)\n",
    "plt.title('Residual plot')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "checking the magnitude of coefficients\n",
    "\n",
    "predictors = x_train.columns\n",
    "\n",
    "coef = Series(lreg.coef_,predictors).sort_values()\n",
    "\n",
    "coef.plot(kind='bar', title='Modal Coefficients')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 503,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW8AAAD4CAYAAAAjKGdbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngIUIbwCw0P8B1qhCtEO9KvcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x15ff98446a0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot(df_train.ac_id, df_train.logerror, 'ro')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 411,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAD3CAYAAAAHQMOGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhQSCIAiC4OX/ARhRDpq0pCmsAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x15fd0762b38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#from pandas.tools.plotting import scatter_matrix\n",
    "#scatter_matrix(df_train[['building_quality_id','yearbuilt']])\n",
    "#plt.show()\n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(df_train.latitude, df_train.longitude, 'ro')\n",
    "plt.show()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}