0

I am trying to find out a solution to Kaggle TFI Dataset.

But while building the model in R using Random Forest I found the following Error:

Error in [.data.frame(data, , all.vars(Terms), drop = FALSE) : undefined columns selected

Code:

library(Boruta)
library(caret)

train<-read.csv("train.csv")
names(train)
View(train)
str(train)
test<-read.csv("test.csv")

n.train<-nrow(train)

test$revenue<-1
myData<-rbind(train,test)
rm(train,test)

myData$Open.Date<-as.POSIXlt("01/01/2015", format="%m/%d/%Y") - 
as.POSIXlt(myData$Open.Date, format="%m/%d/%Y")
myData$Open.Date<-as.numeric(myData$Open.Date/1000)

myData$City<-as.character(myData$City)
myData$City[myData$City.Group=="Other"]<-"Other"
myData$City[myData$City==unique(myData$City)[4]]<-unique(myData$City)[2]
myData$City<-as.factor(myData$City)
myData$City.Group<-NULL

myData$Type<-as.character(myData$Type)
myData$Type[myData$Type=="DT"]<-"IL"
myData$Type[myData$Type=="MB"]<-"FC"
myData$Type<-as.factor(myData$Type)

hist(train$revenue)
hist(log(train$revenue))

myData[,paste("P",1:37,sep="")]<-log(1+myData[,paste("P",1:37,sep="")])

myData$revenue<-log(myData$revenue)
head(myData,2)

important<-Boruta(revenue~.,data=myData[1:n.train,])

important$finalDecision

library(randomForest)
model<-train(myData$revenue[1:n.train]~.,method="rf",
         data=myData[1:n.train,c(important$finalDecision!="Rejected",TRUE)])
BISWAJEET BHATTA
  • 1
  • 1
  • 1
    Welcome to Stack Overflow. Will you please read and incorporate features from [How to make a great R reproducible example?](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example?rq=1)? I looks like your example code above contains a lot of complexity that isn't necessary to reproduce the error. – wibeasley Jun 29 '18 at 19:14
  • check the result of `myData[1:n.train,c(important$finalDecision!="Rejected",TRUE)]` – phiver Jun 30 '18 at 07:27
  • thanks for the response... the code works fine in my console...but, how can i able to apply random forest to create a model for my dataset?? when i am creating a model in random forest using "model<-train(myData$revenue[1:n.train]~.,method="rf", data=myData[1:n.train,c(important$finalDecision!="Rejected",TRUE)])" it is saying "Error in [.data.frame(data, , all.vars(Terms), drop = FALSE) : undefined columns selected" – BISWAJEET BHATTA Jul 01 '18 at 15:56

0 Answers0