** Problem Statement **
This is a retail project where our challenge is to predict whether a retail store should get opened or not based on certain factors such as sales, population,area etc. We have been given two datasets store_train.csv and store_test.csv .We need to use data store_train to build predictive model for response variable ‘store’. store_test data contains all other factors except ‘store’, we need to predict that using the model that we will develop. We will be submitting our predicted values in terms of probability scores. This is a typical classification problem & we will use random forest for model building.
Let’s start Setting our working directory first.
setwd("D:/Edvancer/R Tutorials/R Projects Codes/P2-Retail")Next we need to import both train & test data sets.
s_train=read.csv("store_train.csv",stringsAsFactors = F)
s_test=read.csv("store_test.csv",stringsAsFactors = F)Let us load data wrangling library dplyr so as to glimpse our data.
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionglimpse(s_train)## Observations: 3,338
## Variables: 17
## $ Id <dbl> 2300919770, 5000129575, 2501308470, 603599999, ...
## $ sales0 <int> 848, 925, 924, 924, 1017, 1494, 691, 918, 931, ...
## $ sales1 <int> 588, 717, 616, 646, 730, 1071, 476, 663, 628, 4...
## $ sales2 <int> 666, 780, 739, 683, 735, 1196, 541, 774, 775, 4...
## $ sales3 <int> 1116, 1283, 1154, 1292, 1208, 1861, 861, 1189, ...
## $ sales4 <int> 1133, 1550, 1314, 1297, 1326, 2023, 923, 1477, ...
## $ country <int> 9, 1, 13, 35, 27, 9, 103, 183, 89, 57, 3, 109, ...
## $ State <int> 23, 50, 25, 6, 50, 25, 26, 37, 12, 5, 53, 28, 3...
## $ CouSub <int> 19770, 29575, 8470, 99999, 60100, 37995, 99999,...
## $ countyname <chr> "Hancock County", "Addison County", "Hampden Co...
## $ storecode <chr> "NCNTY23009N23009", "NCNTY50001N50001", "METRO4...
## $ Areaname <chr> "Hancock County, ME", "Addison County, VT", "Sp...
## $ countytownname <chr> "Eastbrook town", "Granville town", "Brimfield ...
## $ population <int> 423, 298, 3609, 34895, 1139, 5136, 67077, 90099...
## $ state_alpha <chr> "ME", "VT", "MA", "CA", "VT", "MA", "MI", "NC",...
## $ store_Type <chr> "Supermarket Type1", "Supermarket Type1", "Supe...
## $ store <int> 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 0, 1,...glimpse(s_test)## Observations: 1,431
## Variables: 16
## $ Id <dbl> 101799999, 101999999, 102199999, 103599999, 103...
## $ sales0 <int> 696, 599, 599, 599, 599, 599, 788, 599, 757, 59...
## $ sales1 <int> 511, 481, 423, 459, 481, 460, 628, 484, 450, 43...
## $ sales2 <int> 514, 500, 475, 462, 505, 463, 663, 505, 572, 46...
## $ sales3 <int> 867, 883, 802, 883, 746, 866, 1084, 746, 950, 8...
## $ sales4 <int> 1034, 894, 1061, 886, 801, 961, 1288, 870, 1079...
## $ country <int> 17, 19, 21, 35, 37, 39, 51, 53, 65, 67, 73, 83,...
## $ State <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,...
## $ CouSub <int> 99999, 99999, 99999, 99999, 99999, 99999, 99999...
## $ countyname <chr> "Chambers County", "Cherokee County", "Chilton ...
## $ storecode <chr> "NCNTY01017N01017", "NCNTY01019N01019", "METRO1...
## $ Areaname <chr> "Chambers County, AL", "Cherokee County, AL", "...
## $ countytownname <chr> "Chambers County", "Cherokee County", "Chilton ...
## $ population <int> 34215, 25989, 43643, 13228, 11539, 37765, 79303...
## $ state_alpha <chr> "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL",...
## $ store_Type <chr> "Supermarket Type1", "Supermarket Type1", "Supe...** Understanding our Data **
Each row represnts characteristic of a single planned store.We can see from above that many categorical data has been coded to mask the data.Here is the interpretation for the columns Id : store id numeric sale figures for 5 types : sales0 sales1 sales2 sales3 sales4
country : categorical :: coded values for country
State : categorical :: coded values for State
CouSub : numeric ::subscription values at county level
countyname : Categorical ::county names
storecode : categorical :: store codes
Areaname : categorical :: name of the area , many times it matches with county name
countytownname : categorical :: county town name
population : numeric :: population of the store area
state_alpha : categorical :: short codes for state
store_Type : categorical :: type of store
store : categorical 1/0 : target indicator var 1=opened 0=not opened
** Data Preparation **
We’ll combine our two datasets so that we do not need to prepare data separately for them. And we’ll also avoid problem of dealing with different columns in different datasets. However before combining them, we’ll need to add response column to test because number of columns need to be same for two datasets to stack vertically.We are also going to add an identifier column ‘data’ which will recognize whether it is from train or test.
s_test$store=NA
s_train$data="train"
s_test$data="test"
s=rbind(s_train,s_test)Let us glimpse our combined data sets s using glimpse & str function.
glimpse(s)## Observations: 4,769
## Variables: 18
## $ Id <dbl> 2300919770, 5000129575, 2501308470, 603599999, ...
## $ sales0 <int> 848, 925, 924, 924, 1017, 1494, 691, 918, 931, ...
## $ sales1 <int> 588, 717, 616, 646, 730, 1071, 476, 663, 628, 4...
## $ sales2 <int> 666, 780, 739, 683, 735, 1196, 541, 774, 775, 4...
## $ sales3 <int> 1116, 1283, 1154, 1292, 1208, 1861, 861, 1189, ...
## $ sales4 <int> 1133, 1550, 1314, 1297, 1326, 2023, 923, 1477, ...
## $ country <int> 9, 1, 13, 35, 27, 9, 103, 183, 89, 57, 3, 109, ...
## $ State <int> 23, 50, 25, 6, 50, 25, 26, 37, 12, 5, 53, 28, 3...
## $ CouSub <int> 19770, 29575, 8470, 99999, 60100, 37995, 99999,...
## $ countyname <chr> "Hancock County", "Addison County", "Hampden Co...
## $ storecode <chr> "NCNTY23009N23009", "NCNTY50001N50001", "METRO4...
## $ Areaname <chr> "Hancock County, ME", "Addison County, VT", "Sp...
## $ countytownname <chr> "Eastbrook town", "Granville town", "Brimfield ...
## $ population <int> 423, 298, 3609, 34895, 1139, 5136, 67077, 90099...
## $ state_alpha <chr> "ME", "VT", "MA", "CA", "VT", "MA", "MI", "NC",...
## $ store_Type <chr> "Supermarket Type1", "Supermarket Type1", "Supe...
## $ store <int> 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 0, 1,...
## $ data <chr> "train", "train", "train", "train", "train", "t...str(s)## 'data.frame': 4769 obs. of 18 variables:
## $ Id : num 2.30e+09 5.00e+09 2.50e+09 6.04e+08 5.00e+09 ...
## $ sales0 : int 848 925 924 924 1017 1494 691 918 931 584 ...
## $ sales1 : int 588 717 616 646 730 1071 476 663 628 455 ...
## $ sales2 : int 666 780 739 683 735 1196 541 774 775 477 ...
## $ sales3 : int 1116 1283 1154 1292 1208 1861 861 1189 1228 727 ...
## $ sales4 : int 1133 1550 1314 1297 1326 2023 923 1477 1502 949 ...
## $ country : int 9 1 13 35 27 9 103 183 89 57 ...
## $ State : int 23 50 25 6 50 25 26 37 12 5 ...
## $ CouSub : int 19770 29575 8470 99999 60100 37995 99999 99999 99999 99999 ...
## $ countyname : chr "Hancock County" "Addison County" "Hampden County" "Lassen County" ...
## $ storecode : chr "NCNTY23009N23009" "NCNTY50001N50001" "METRO44140M44140" "NCNTY06035N06035" ...
## $ Areaname : chr "Hancock County, ME" "Addison County, VT" "Springfield, MA HUD Metro FMR Area" "Lassen County, CA" ...
## $ countytownname: chr "Eastbrook town" "Granville town" "Brimfield town" "Lassen County" ...
## $ population : int 423 298 3609 34895 1139 5136 67077 900993 73314 22609 ...
## $ state_alpha : chr "ME" "VT" "MA" "CA" ...
## $ store_Type : chr "Supermarket Type1" "Supermarket Type1" "Supermarket Type1" "Supermarket Type3" ...
## $ store : int 0 0 1 0 0 0 0 1 1 0 ...
## $ data : chr "train" "train" "train" "train" ...Many categorical data like ‘country’ & ‘State’ has already been coded to mask the data. We can see the same using frequency table as shown below.
table(s$country)##
## 1 3 5 6 7 9 10 11 12 13 14 15 16 17 19 20 21 23
## 133 219 163 1 168 213 2 165 1 131 1 145 1 165 144 2 121 100
## 25 27 28 29 30 31 33 35 36 37 39 41 43 45 47 49 50 51
## 104 152 2 90 1 70 41 40 1 40 39 39 39 39 38 37 1 37
## 53 54 55 56 57 59 60 61 63 65 67 68 69 70 71 73 75 77
## 37 1 36 1 37 36 1 36 35 35 35 1 35 1 35 34 34 34
## 78 79 81 83 85 86 87 89 90 91 93 95 97 99 100 101 103 105
## 1 33 33 33 33 1 33 32 1 32 31 31 31 31 1 31 31 32
## 107 109 110 111 113 115 117 119 121 122 123 125 127 129 130 131 133 135
## 30 30 1 29 28 28 27 27 27 1 26 26 25 23 1 23 23 21
## 137 139 141 143 145 147 149 150 151 153 155 157 159 161 163 164 165 167
## 21 20 20 19 19 19 19 1 17 18 16 16 16 16 16 1 15 14
## 169 170 171 173 175 177 179 180 181 183 185 186 187 188 189 191 193 195
## 14 1 14 14 13 12 12 1 12 12 12 1 10 1 9 8 8 10
## 197 198 199 201 203 205 207 209 211 213 215 217 219 220 221 223 225 227
## 9 1 8 6 5 5 5 5 4 4 4 4 4 1 4 4 4 4
## 229 230 231 233 235 237 239 240 241 243 245 247 249 251 253 255 257 259
## 4 1 3 3 3 3 3 1 2 2 2 2 2 2 2 2 2 2
## 261 263 265 267 269 270 271 273 275 277 279 281 282 283 285 287 289 290
## 3 2 2 2 2 1 2 2 3 2 2 2 1 2 2 2 2 1
## 291 293 295 297 299 301 303 305 307 309 311 313 315 317 319 321 323 325
## 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1
## 327 329 331 333 335 337 339 341 343 345 347 349 351 353 355 357 359 361
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 363 365 367 369 371 373 375 377 379 381 383 385 387 389 391 393 395 397
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 399 401 403 405 407 409 411 413 415 417 419 421 423 425 427 429 431 433
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 435 437 439 441 443 445 447 449 451 453 455 457 459 461 463 465 467 469
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 471 473 475 477 479 481 483 485 487 489 491 493 495 497 499 501 503 505
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 507 510 515 520 530 540 550 560 570 580 590 595 600 610 620 630 640 650
## 1 4 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 660 670 678 680 683 685 690 700 710 720 730 735 740 750 760 770 775 790
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 800 810 820 830 840
## 1 1 1 1 1table(s$State)##
## 1 2 4 5 6 8 9 10 11 12 13 15 16 17 18 19 20 21
## 67 29 15 75 58 64 169 3 1 67 159 5 44 102 92 99 105 120
## 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
## 64 535 25 351 83 87 82 116 56 93 17 259 21 33 62 100 53 88
## 40 41 42 44 45 46 47 48 49 50 51 53 54 55 56 66 72 78
## 77 36 67 39 46 66 95 254 29 255 135 39 55 72 23 1 78 3Since we will be using random forest we need to convert data type of response (which is store in this case) to factor type using function as.factor. This is how randomforest differentiates from regression & classification.If we need to build a regression model then response variable should be kept numeric else factor for classification.
s$store=as.factor(s$store)Let us see if store column has been changed to factor type or not using glimpse function again from dplyr package.
glimpse(s)## Observations: 4,769
## Variables: 18
## $ Id <dbl> 2300919770, 5000129575, 2501308470, 603599999, ...
## $ sales0 <int> 848, 925, 924, 924, 1017, 1494, 691, 918, 931, ...
## $ sales1 <int> 588, 717, 616, 646, 730, 1071, 476, 663, 628, 4...
## $ sales2 <int> 666, 780, 739, 683, 735, 1196, 541, 774, 775, 4...
## $ sales3 <int> 1116, 1283, 1154, 1292, 1208, 1861, 861, 1189, ...
## $ sales4 <int> 1133, 1550, 1314, 1297, 1326, 2023, 923, 1477, ...
## $ country <int> 9, 1, 13, 35, 27, 9, 103, 183, 89, 57, 3, 109, ...
## $ State <int> 23, 50, 25, 6, 50, 25, 26, 37, 12, 5, 53, 28, 3...
## $ CouSub <int> 19770, 29575, 8470, 99999, 60100, 37995, 99999,...
## $ countyname <chr> "Hancock County", "Addison County", "Hampden Co...
## $ storecode <chr> "NCNTY23009N23009", "NCNTY50001N50001", "METRO4...
## $ Areaname <chr> "Hancock County, ME", "Addison County, VT", "Sp...
## $ countytownname <chr> "Eastbrook town", "Granville town", "Brimfield ...
## $ population <int> 423, 298, 3609, 34895, 1139, 5136, 67077, 90099...
## $ state_alpha <chr> "ME", "VT", "MA", "CA", "VT", "MA", "MI", "NC",...
## $ store_Type <chr> "Supermarket Type1", "Supermarket Type1", "Supe...
## $ store <fct> 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 0, 1,...
## $ data <chr> "train", "train", "train", "train", "train", "t...Next we will convert all categorical variables to dummies. We will write a function which will take care of that instead of converting them one by one.
** Writing a Dummy Creation Function **
CreateDummies=function(data,var,freq_cutoff=0){
t=table(data[,var])
t=t[t>freq_cutoff]
t=sort(t)
categories=names(t)[-1]
for( cat in categories){
name=paste(var,cat,sep="_")
name=gsub(" ","",name)
name=gsub("-","_",name)
name=gsub("\\?","Q",name)
name=gsub("<","LT_",name)
name=gsub("\\+","",name)
name=gsub("\\/","_",name)
name=gsub(">","GT_",name)
name=gsub("=","EQ_",name)
name=gsub(",","",name)
data[,name]=as.numeric(data[,var]==cat)
}
data[,var]=NULL
return(data)
}Let me explain the function ‘CreateDummies’ we just created
t=table(data[,var]) this bit creates a frequency table for the given categorical column. t here is now simply a table which contains names as categories of the categorical variable and their frequency in the data.
t=t[t>freq_cutoff] this line of code removes those categories from the table which have frequencies below the frequency cutoff. ( this is a subjective choice)
‘t=sort(t)’ this line simple sorts the remaining table in ascending order
categories=names(t)[-1] since we sorted the table in ascending manner in the previous line, first category here has least count. In this line of code we are taking out the category names except the first one ( which has least count), thus making n-1 dummies from the remaining categories.
name=paste(var,cat,sep=“_“) all the dummy vars that we intend to make, need to have some name. this line of code creates that name by concatenating variable name with category name with an _.
name=gsub(" “,”“,name) subsequent lines like these using gsub are essentially cleaning up the name. Since we dont have any control over what the categories can be, we are removing special characters and spaces in the code in an automated fashion.
data[,name]=as.numeric(data[,var]==cat) once we have a cleaned up name, this line creates the dummy var for that particular category.
data[,var]=NULL once we are done creating dummies for the variable using for loop. Variable is removed from the data in this line.
Let us have a look at our categorical variables by writing following lines of codes
names(s)[sapply(s,function(x) is.character(x))]## [1] "countyname" "storecode" "Areaname" "countytownname"
## [5] "state_alpha" "store_Type" "data"Now we will check for High-Cardinality in the categorical variables i.e we will check for variables with many distinct values. We will discard those variables from our modelling. Because including these attributes by standard dummy encoding increases the dimensionality of the data to such an extent that either the classification technique is unable to process them or if one would use some regularized linear technique that is able to cope with huge dimensions, it leads to a model with thousands or even millions of features, thereby losing the often required comprehensibility aspect.
length(unique(s$countyname))## [1] 1962length(unique(s$storecode))## [1] 2572length(unique(s$Areaname))## [1] 2572length(unique(s$countytownname))## [1] 3176length(unique(s$state_alpha))## [1] 54length(unique(s$store_Type))## [1] 4We will ignore columns or variables like countyname,storecode,Areaname,countytownname for their High-Cardinality. Further we will ignore data column for obvious reason.
s=s %>% select(-countyname,-storecode,-Areaname,-countytownname)Above codes will discard those four variables & we are left with 14 variables now. Next Let us make dummies for the rest of columns - state_alpha & store_Type.
cat_cols=c("state_alpha","store_Type")
for(cat in cat_cols){
s=CreateDummies(s,cat,100)
}This will increase our columns to 26 as we can glimpse the same as shown below.
glimpse(s)## Observations: 4,769
## Variables: 26
## $ Id <dbl> 2300919770, 5000129575, 2501308470...
## $ sales0 <int> 848, 925, 924, 924, 1017, 1494, 69...
## $ sales1 <int> 588, 717, 616, 646, 730, 1071, 476...
## $ sales2 <int> 666, 780, 739, 683, 735, 1196, 541...
## $ sales3 <int> 1116, 1283, 1154, 1292, 1208, 1861...
## $ sales4 <int> 1133, 1550, 1314, 1297, 1326, 2023...
## $ country <int> 9, 1, 13, 35, 27, 9, 103, 183, 89,...
## $ State <int> 23, 50, 25, 6, 50, 25, 26, 37, 12,...
## $ CouSub <int> 19770, 29575, 8470, 99999, 60100, ...
## $ population <int> 423, 298, 3609, 34895, 1139, 5136,...
## $ store <fct> 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1...
## $ data <chr> "train", "train", "train", "train"...
## $ state_alpha_KS <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_MO <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_KY <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_VA <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_GA <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_CT <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_TX <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_VT <dbl> 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_NH <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ state_alpha_MA <dbl> 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0...
## $ state_alpha_ME <dbl> 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ store_Type_SupermarketType3 <dbl> 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0...
## $ store_Type_GroceryStore <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0...
## $ store_Type_SupermarketType1 <dbl> 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 1, 1...Let us see if there is any missing values in our data. We will use lapply function which will give output in list format as shown below.
lapply(s,function(x) sum(is.na(x)))## $Id
## [1] 0
##
## $sales0
## [1] 0
##
## $sales1
## [1] 0
##
## $sales2
## [1] 0
##
## $sales3
## [1] 0
##
## $sales4
## [1] 0
##
## $country
## [1] 1
##
## $State
## [1] 0
##
## $CouSub
## [1] 0
##
## $population
## [1] 2
##
## $store
## [1] 1431
##
## $data
## [1] 0
##
## $state_alpha_KS
## [1] 0
##
## $state_alpha_MO
## [1] 0
##
## $state_alpha_KY
## [1] 0
##
## $state_alpha_VA
## [1] 0
##
## $state_alpha_GA
## [1] 0
##
## $state_alpha_CT
## [1] 0
##
## $state_alpha_TX
## [1] 0
##
## $state_alpha_VT
## [1] 0
##
## $state_alpha_NH
## [1] 0
##
## $state_alpha_MA
## [1] 0
##
## $state_alpha_ME
## [1] 0
##
## $store_Type_SupermarketType3
## [1] 0
##
## $store_Type_GroceryStore
## [1] 0
##
## $store_Type_SupermarketType1
## [1] 0From above we can see that We do have missing values in columns like country, population & store. Next we impute those missing values with the mean of train data as shown below.
for(col in names(s)){
if(sum(is.na(s[,col]))>0 & !(col %in% c("data","store"))){
s[is.na(s[,col]),col]=mean(s[s$data=='train',col],na.rm=T)
}
}We can always cross check if those NAs has been replaced with mean or not by using lapply function again.
lapply(s,function(x) sum(is.na(x)))## $Id
## [1] 0
##
## $sales0
## [1] 0
##
## $sales1
## [1] 0
##
## $sales2
## [1] 0
##
## $sales3
## [1] 0
##
## $sales4
## [1] 0
##
## $country
## [1] 0
##
## $State
## [1] 0
##
## $CouSub
## [1] 0
##
## $population
## [1] 0
##
## $store
## [1] 1431
##
## $data
## [1] 0
##
## $state_alpha_KS
## [1] 0
##
## $state_alpha_MO
## [1] 0
##
## $state_alpha_KY
## [1] 0
##
## $state_alpha_VA
## [1] 0
##
## $state_alpha_GA
## [1] 0
##
## $state_alpha_CT
## [1] 0
##
## $state_alpha_TX
## [1] 0
##
## $state_alpha_VT
## [1] 0
##
## $state_alpha_NH
## [1] 0
##
## $state_alpha_MA
## [1] 0
##
## $state_alpha_ME
## [1] 0
##
## $store_Type_SupermarketType3
## [1] 0
##
## $store_Type_GroceryStore
## [1] 0
##
## $store_Type_SupermarketType1
## [1] 0Now we are done with data preparation , lets separate the data next.
s_train=s %>% filter(data=="train") %>% select(-data)
s_test=s %>% filter(data=="test") %>% select(-data,-store)Next we will break our train data into 2 parts. We will build model on one part & check its performance on the other.
set.seed(2)
s=sample(1:nrow(s_train),0.8*nrow(s_train))
s_train1=s_train[s,]
s_train2=s_train[-s,]** Model Building **
Let us load the package randomForest first
library(randomForest)## randomForest 4.6-14## Type rfNews() to see new features/changes/bug fixes.##
## Attaching package: 'randomForest'## The following object is masked from 'package:dplyr':
##
## combineNext we will build our model with 5 variables randomly subsetted at each node i.e mtry & let just say we want to grow 100 such trees.
model_rf=randomForest(store~.-Id,data=s_train1,mtry=5,ntree=100)Let us see what does this model_rf represent
model_rf##
## Call:
## randomForest(formula = store ~ . - Id, data = s_train1, mtry = 5, ntree = 100)
## Type of random forest: classification
## Number of trees: 100
## No. of variables tried at each split: 5
##
## OOB estimate of error rate: 23.41%
## Confusion matrix:
## 0 1 class.error
## 0 1273 243 0.1602902
## 1 382 772 0.3310225It is clear from above that OOB estimate of error has come out to be 23.22 % which is decent. Also from the confusion matrix we can conclude that our model has correctly predicted for 1277 stores as ‘not opened’ & for 773 stores correctly predicted as ‘opened’. The diagonal values 381 & 239 are giving justification as to why our OOB error is around 23%.
** Model Validation **
Lets see performance of this model on the validation data s_train2 that we kept aside.
val.score=predict(model_rf,newdata=s_train2,type='response')Again we need to check the accuracy using confusionMatrix from caret package. What we will get is an accuracy of 78% which seems to be a fair model.
library(caret)## Loading required package: lattice## Loading required package: ggplot2##
## Attaching package: 'ggplot2'## The following object is masked from 'package:randomForest':
##
## marginconfusionMatrix(val.score,s_train2$store)## Confusion Matrix and Statistics
##
## Reference
## Prediction 0 1
## 0 311 91
## 1 48 218
##
## Accuracy : 0.7919
## 95% CI : (0.7591, 0.8221)
## No Information Rate : 0.5374
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.5774
##
## Mcnemar's Test P-Value : 0.0003675
##
## Sensitivity : 0.8663
## Specificity : 0.7055
## Pos Pred Value : 0.7736
## Neg Pred Value : 0.8195
## Prevalence : 0.5374
## Detection Rate : 0.4656
## Detection Prevalence : 0.6018
## Balanced Accuracy : 0.7859
##
## 'Positive' Class : 0
## Now let us calculate probability score for our validation data set s_train2.
val.prob_score=predict(model_rf,newdata=s_train2,type='prob')In order to check the performance of our model let us calculate its auc score. For that we need to first import a package named ‘pROC’.
library(pROC)## Type 'citation("pROC")' for a citation.##
## Attaching package: 'pROC'## The following objects are masked from 'package:stats':
##
## cov, smooth, varauc_score=auc(roc(s_train2$store,val.prob_score[,1]))auc_score## Area under the curve: 0.8236From above it is clear that the auc score or the tentative score performance of our model is going to be around 0.82 which is decent enough.
We can also plot our auc
plot(roc(s_train2$store,val.prob_score[,1]))
Next we will build the random forest model on the entire training data set ‘s_train’ & predict the same on test data set ‘s_test’
model_rf_final=randomForest(store~.-Id,data=s_train,mtry=5,ntree=100)model_rf_final##
## Call:
## randomForest(formula = store ~ . - Id, data = s_train, mtry = 5, ntree = 100)
## Type of random forest: classification
## Number of trees: 100
## No. of variables tried at each split: 5
##
## OOB estimate of error rate: 22.47%
## Confusion matrix:
## 0 1 class.error
## 0 1583 292 0.1557333
## 1 458 1005 0.3130554We will now use this model to predict probability score for test data .
test.score=predict(model_rf_final,newdata = s_test,type='prob')[,1]To see what does my test.score contains
test.score## 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
## 0.72 0.94 0.66 0.94 0.82 0.78 0.32 0.82 0.86 0.85 0.32 0.42 0.88 0.19 0.70
## 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
## 0.61 0.86 0.75 0.52 0.61 0.53 0.58 0.75 0.34 0.63 0.40 0.34 0.28 0.90 0.71
## 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
## 0.96 0.58 0.88 0.71 0.72 0.78 0.79 0.37 0.93 0.59 0.93 0.67 0.99 0.96 0.86
## 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
## 0.91 0.87 0.15 0.66 0.68 0.38 0.08 0.08 0.56 0.35 0.24 0.73 0.41 0.89 0.28
## 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
## 0.61 0.82 0.80 0.70 0.37 0.87 0.38 0.77 0.65 0.91 0.71 0.28 0.78 0.14 0.04
## 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90
## 0.01 0.03 0.01 0.11 0.06 0.13 0.05 0.13 0.25 0.06 0.05 0.34 0.45 0.56 0.71
## 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
## 0.80 0.68 0.64 0.30 0.11 0.19 0.49 0.23 0.16 0.05 0.66 0.14 0.42 0.11 0.19
## 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
## 0.01 0.17 0.22 0.07 0.15 0.22 0.09 0.09 0.72 0.76 0.78 0.27 0.29 0.84 0.05
## 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135
## 0.34 0.92 0.56 0.25 0.24 0.38 0.40 0.17 0.77 0.93 0.17 0.73 0.13 0.05 0.28
## 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
## 0.44 0.66 0.85 0.73 0.52 0.83 0.58 0.34 0.17 0.79 0.71 0.08 0.09 0.83 0.31
## 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
## 0.59 0.23 0.90 0.69 0.47 0.89 0.79 0.08 0.44 0.80 0.86 0.08 0.92 0.22 0.74
## 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
## 0.90 0.91 0.98 0.11 0.91 0.78 0.10 0.69 0.71 0.47 0.67 0.70 0.92 0.97 0.79
## 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
## 0.37 0.94 0.48 0.81 0.87 0.89 0.92 0.55 0.83 0.21 0.89 0.48 0.78 0.37 0.80
## 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
## 0.73 0.39 0.91 0.64 0.88 0.36 0.76 0.90 0.30 0.17 0.03 0.78 0.49 0.76 0.42
## 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
## 0.18 0.22 0.86 0.87 0.54 0.81 0.82 0.73 0.50 0.49 0.80 0.74 0.51 0.67 0.18
## 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
## 0.41 0.24 0.05 0.81 0.79 0.08 0.60 0.14 0.27 0.48 0.47 0.74 0.66 0.66 0.76
## 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255
## 0.14 0.62 0.73 0.90 0.18 0.88 0.34 0.75 0.89 0.51 0.31 0.84 0.89 0.66 0.97
## 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
## 0.90 0.74 0.91 0.75 0.83 0.79 0.25 0.85 0.77 0.95 0.67 0.74 0.82 0.69 0.78
## 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
## 0.70 0.72 0.72 0.54 0.56 0.93 0.85 0.25 0.21 0.83 0.65 0.71 0.75 0.80 0.83
## 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
## 0.39 0.69 0.51 0.24 0.85 0.76 0.93 0.57 0.87 0.65 0.88 0.77 0.24 0.51 0.89
## 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
## 0.52 0.23 0.80 0.29 0.09 0.45 0.91 0.19 0.35 0.87 0.78 0.78 0.89 0.36 0.88
## 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330
## 0.69 0.83 0.04 0.66 0.13 0.65 0.88 0.19 0.63 0.79 0.46 0.08 0.82 0.08 0.20
## 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345
## 0.89 0.18 0.15 0.41 0.64 0.20 0.67 0.22 0.53 0.81 0.45 0.11 0.19 0.21 0.84
## 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
## 0.91 0.98 0.96 0.95 0.87 0.99 0.97 0.91 0.95 0.97 0.92 0.92 0.96 0.90 0.97
## 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
## 0.96 0.83 0.15 0.01 0.02 0.03 0.00 0.78 0.74 0.88 0.84 0.91 0.97 0.92 0.84
## 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
## 0.97 0.92 0.89 0.89 0.73 0.94 0.88 0.81 0.92 0.88 0.83 0.63 0.51 0.93 0.86
## 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
## 0.84 0.88 0.96 0.98 0.93 0.90 0.92 0.99 0.67 0.57 0.94 0.99 0.88 0.97 0.95
## 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420
## 0.97 0.86 0.91 0.85 0.98 0.91 0.98 0.96 0.80 0.97 0.96 0.10 0.04 0.04 0.05
## 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
## 0.14 0.09 0.04 0.14 0.15 0.11 0.14 0.02 0.09 0.06 0.12 0.10 0.13 0.08 0.10
## 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450
## 0.02 0.79 0.77 0.87 0.95 0.97 0.96 0.99 0.41 0.40 0.55 0.86 0.98 0.98 0.88
## 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
## 0.81 0.91 0.92 0.85 0.88 0.92 0.98 0.97 1.00 0.97 1.00 0.90 0.95 0.84 0.87
## 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
## 0.71 0.70 0.95 0.95 0.93 0.91 0.97 0.98 0.97 0.99 0.96 0.94 0.07 0.14 0.09
## 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
## 0.11 0.16 0.68 0.06 0.58 0.13 0.02 0.12 0.64 0.17 0.22 0.02 0.12 0.17 0.10
## 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510
## 0.07 0.00 0.03 0.03 0.02 0.51 0.67 0.71 0.05 0.10 0.48 0.27 0.09 0.17 0.36
## 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525
## 0.04 0.25 0.02 0.04 0.09 0.20 0.05 0.00 0.06 0.08 0.24 0.12 0.04 0.13 0.00
## 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540
## 0.08 0.25 0.02 0.27 0.27 0.16 0.01 0.00 0.04 0.02 0.10 0.00 0.01 0.02 0.03
## 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555
## 0.00 0.04 0.00 0.02 0.15 0.15 0.15 0.31 0.15 0.20 0.13 0.14 0.16 0.03 0.02
## 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570
## 0.24 0.07 0.26 0.13 0.06 0.26 0.06 0.39 0.06 0.22 0.57 0.77 0.73 0.58 0.79
## 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
## 0.76 0.89 0.91 0.49 0.91 0.73 0.86 0.83 0.60 0.07 0.23 0.73 0.40 0.68 0.43
## 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
## 0.92 0.95 0.78 0.49 0.93 0.90 0.90 0.68 0.70 0.81 0.72 0.29 0.88 0.15 0.88
## 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615
## 0.87 0.10 0.79 0.92 0.76 0.89 0.40 0.84 0.90 0.82 0.75 0.81 0.88 0.86 0.71
## 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630
## 0.54 0.71 0.69 0.85 0.83 0.31 0.77 0.90 0.67 0.69 0.77 0.75 0.89 0.83 0.69
## 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645
## 0.30 0.79 0.53 0.75 0.87 0.66 0.92 0.88 0.77 0.03 0.41 0.75 0.77 0.75 0.99
## 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660
## 0.91 0.78 0.84 0.76 0.60 0.77 0.84 0.72 0.75 0.74 0.56 0.98 0.58 0.84 0.65
## 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675
## 0.87 0.81 0.57 0.59 0.97 0.87 0.95 0.85 0.86 0.80 0.42 0.87 0.76 0.86 0.53
## 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690
## 0.73 0.24 0.97 0.98 0.86 0.85 0.59 0.99 0.95 0.70 0.77 0.98 0.86 0.17 1.00
## 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705
## 0.90 0.44 0.69 0.56 0.57 0.78 0.91 0.67 0.89 0.85 0.82 0.99 0.97 0.99 0.97
## 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720
## 0.87 0.61 0.98 0.96 0.99 0.69 0.93 0.60 0.86 0.93 0.97 0.66 0.94 0.98 0.92
## 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735
## 0.93 0.98 0.99 0.97 0.71 1.00 0.78 0.89 0.99 1.00 0.92 0.85 0.90 0.07 0.19
## 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750
## 0.08 0.19 0.24 0.13 0.13 0.64 0.92 0.89 0.92 0.16 0.25 0.06 0.20 0.14 0.09
## 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765
## 0.10 0.09 0.42 0.13 0.32 0.19 0.05 0.07 0.04 0.95 0.84 0.96 0.96 0.97 0.97
## 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
## 0.96 0.96 0.92 0.15 0.40 0.02 0.07 0.16 0.12 0.06 0.14 0.13 0.10 0.08 0.31
## 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
## 0.83 0.73 0.85 0.74 0.75 0.48 0.87 0.85 0.52 0.37 0.70 0.51 0.90 0.20 0.08
## 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810
## 0.05 0.19 0.22 0.41 0.44 0.17 0.86 0.77 0.81 0.58 0.55 0.53 0.43 0.40 0.58
## 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825
## 0.81 0.68 0.64 0.10 0.77 0.34 0.82 0.19 0.57 0.82 0.49 0.49 0.80 0.74 0.97
## 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840
## 0.79 0.96 0.73 0.82 0.87 0.44 0.69 0.93 0.42 0.86 0.91 0.88 0.96 0.69 0.29
## 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855
## 0.36 0.62 0.40 0.64 0.25 0.76 0.38 0.69 0.28 0.13 0.59 0.39 0.50 0.79 0.68
## 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
## 0.18 0.19 0.88 0.21 0.29 0.76 0.83 0.71 0.80 0.80 0.70 0.82 0.98 0.18 0.40
## 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885
## 0.68 0.89 0.13 0.83 0.77 0.61 0.53 0.75 0.78 0.87 0.42 0.31 0.48 0.86 0.10
## 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900
## 0.42 0.55 0.40 0.54 0.63 0.39 0.25 0.29 0.45 0.90 0.79 0.44 0.75 0.55 0.38
## 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915
## 0.39 0.84 0.56 0.33 0.35 0.82 0.04 0.45 0.96 0.68 0.21 0.56 0.10 0.22 0.14
## 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930
## 0.09 0.41 0.26 0.33 0.68 0.80 0.20 0.63 0.79 0.78 0.56 0.30 0.40 0.59 0.68
## 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945
## 0.72 0.69 0.72 0.98 0.72 0.78 0.58 0.95 0.67 0.96 0.84 0.87 0.21 0.83 0.83
## 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960
## 0.82 0.81 0.55 0.73 0.80 0.69 0.85 0.26 0.59 0.25 0.32 0.77 0.55 0.48 0.65
## 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975
## 0.89 0.70 0.77 0.62 0.79 0.83 0.63 0.21 0.82 0.91 0.61 0.59 0.45 0.43 0.26
## 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990
## 0.73 0.69 0.26 0.55 0.46 0.33 0.58 0.30 0.35 0.18 0.50 0.63 0.79 0.26 0.80
## 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005
## 0.49 0.57 0.52 0.87 0.91 0.93 0.78 0.72 0.73 0.57 0.04 0.70 0.30 0.72 0.81
## 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020
## 0.49 0.66 0.06 0.93 0.86 0.82 0.37 0.52 0.69 0.91 0.84 0.86 0.19 0.90 0.76
## 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
## 0.80 0.80 0.11 0.70 0.93 0.09 0.63 0.68 0.18 0.56 0.68 0.69 0.84 0.90 0.66
## 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
## 0.05 0.70 0.86 0.76 0.07 0.65 0.79 0.24 0.70 0.51 0.67 0.77 0.73 0.92 0.83
## 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
## 0.79 0.57 0.29 0.41 0.76 0.55 0.75 0.72 0.86 0.83 0.73 0.80 0.89 0.97 0.98
## 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
## 0.97 0.97 0.99 0.99 0.99 0.96 0.93 1.00 0.98 0.99 0.26 0.22 0.15 0.24 0.34
## 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
## 0.98 0.98 1.00 1.00 0.52 0.18 0.55 0.17 0.13 0.08 0.10 0.88 0.78 0.72 0.90
## 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110
## 0.91 0.89 0.85 0.89 0.99 0.97 0.99 0.98 0.83 0.99 1.00 0.90 0.91 0.94 0.96
## 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
## 0.90 0.83 0.99 0.82 0.78 0.92 0.92 0.70 0.60 0.95 0.82 0.54 0.65 0.64 0.46
## 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140
## 0.16 0.75 0.32 0.02 0.77 0.79 0.29 0.82 0.19 0.58 0.11 0.12 0.15 0.03 0.41
## 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
## 0.93 0.23 0.34 0.81 0.14 0.21 0.56 0.50 0.76 0.15 0.66 0.42 0.55 0.38 0.58
## 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170
## 0.35 0.61 0.66 0.34 0.66 0.79 0.77 0.91 0.22 0.43 0.80 0.83 0.71 0.55 0.90
## 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
## 0.86 0.81 0.33 0.65 0.82 0.88 0.45 0.57 0.82 0.35 0.83 0.92 0.56 0.29 0.76
## 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200
## 0.78 0.40 0.32 0.83 0.83 0.66 0.62 0.51 0.90 0.68 0.78 0.17 0.04 0.03 0.08
## 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
## 0.10 0.03 0.21 0.73 0.01 0.12 0.07 0.03 0.15 0.16 0.06 0.03 0.26 0.31 0.09
## 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
## 0.08 0.10 0.00 0.45 0.76 0.90 0.91 0.84 0.23 0.64 0.06 0.26 0.85 0.71 0.34
## 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245
## 0.17 0.40 0.78 0.94 0.75 0.09 0.33 0.31 0.05 0.96 0.98 0.87 0.91 0.50 0.97
## 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260
## 0.94 0.34 0.89 0.64 0.94 0.94 0.43 0.21 0.74 0.46 0.10 0.93 0.80 0.47 0.61
## 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275
## 0.07 0.86 0.91 0.87 0.11 0.91 0.73 1.00 0.04 0.00 0.23 0.78 0.28 0.84 0.75
## 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
## 0.01 0.88 0.65 0.98 0.78 0.80 0.86 0.76 0.88 0.97 0.61 0.24 0.94 1.00 0.98
## 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305
## 0.55 0.94 0.15 0.25 0.84 0.73 0.48 0.78 0.73 0.49 0.66 0.91 0.21 0.79 0.90
## 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
## 0.71 0.12 0.15 0.15 0.99 0.67 0.07 0.40 0.14 0.15 0.45 0.13 0.97 0.19 0.92
## 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
## 0.87 0.10 0.04 0.92 0.93 0.45 0.60 0.98 0.76 0.11 0.78 0.15 0.14 0.95 0.78
## 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
## 0.84 0.10 0.90 0.03 0.80 0.94 0.40 0.03 0.00 0.76 0.85 0.89 0.81 0.85 0.44
## 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
## 0.20 0.68 0.04 0.02 0.64 0.61 0.55 0.82 0.89 0.15 0.72 0.85 0.98 0.89 0.91
## 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380
## 0.01 0.82 0.81 0.54 0.84 0.63 0.59 0.57 0.96 0.90 0.40 0.10 0.63 0.06 0.58
## 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395
## 0.99 0.12 0.59 0.76 0.10 0.01 0.65 0.08 0.98 0.28 0.63 0.83 0.36 0.82 0.13
## 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410
## 0.17 0.86 0.73 0.97 0.75 0.19 0.77 0.59 0.13 0.07 0.06 0.63 0.98 0.87 0.22
## 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425
## 0.21 0.92 0.98 0.78 0.31 0.05 0.99 0.09 0.54 0.96 0.18 0.30 0.45 0.96 0.90
## 1426 1427 1428 1429 1430 1431
## 0.73 0.84 0.63 0.96 0.96 1.00** Variable Importance **
We will run below codes to find out the importance of variable. Higher the mean decrease ginni for any variable better is the variable for prediction. So population is the most important variable.
d=importance(model_rf_final)
d=as.data.frame(d)
d$VariableNames=rownames(d)
d %>% arrange(desc(MeanDecreaseGini))## MeanDecreaseGini VariableNames
## 1 258.626715 population
## 2 167.133394 sales4
## 3 163.245043 sales0
## 4 161.830217 sales3
## 5 147.395128 sales2
## 6 137.757549 sales1
## 7 123.945385 country
## 8 110.103799 State
## 9 66.276194 CouSub
## 10 26.698305 state_alpha_MA
## 11 24.843971 store_Type_SupermarketType1
## 12 20.072403 state_alpha_VT
## 13 15.871697 store_Type_GroceryStore
## 14 14.412103 store_Type_SupermarketType3
## 15 7.264191 state_alpha_NH
## 16 7.187766 state_alpha_ME
## 17 5.459227 state_alpha_GA
## 18 5.017612 state_alpha_TX
## 19 4.584571 state_alpha_VA
## 20 4.377301 state_alpha_KS
## 21 4.030374 state_alpha_MO
## 22 3.389849 state_alpha_KY
## 23 2.836885 state_alpha_CTUpon plotting we get a plot like this.
varImpPlot(model_rf_final)