Different RF predictions when pasting code from different sources - r

Question

I'm playing around with the German Credit dataset from the "caret" package.
First, I build a very simple model:
library(caret)
library(randomForest)
library(pmml)
data(GermanCredit)
GermanCredit <- GermanCredit[, -nearZeroVar(GermanCredit)]
GermanCredit$CheckingAccountStatus.lt.0 <- NULL
GermanCredit$SavingsAccountBonds.lt.100 <- NULL
GermanCredit$EmploymentDuration.lt.1 <- NULL
GermanCredit$EmploymentDuration.Unemployed <- NULL
GermanCredit$Personal.Male.Married.Widowed <- NULL
GermanCredit$Property.Unknown <- NULL
GermanCredit$Housing.ForFree <- NULL
set.seed(100)
inTrain <- createDataPartition(GermanCredit$Class, p = .8)[[1]]
GermanCreditTrain <- GermanCredit[ inTrain, ]
GermanCreditTest <- GermanCredit[-inTrain, ]
set.seed(1056)
credit.rf <- randomForest(Class~., data = GermanCreditTrain, ntree = 500)
Now, if I predict the outcome Class on the test set, and do this several times, and then compare the results:
credit.pred1 <- predict(credit.rf, GermanCreditTest)
credit.pred2 <- predict(credit.rf, GermanCreditTest)
credit.pred3 <- predict(credit.rf, GermanCreditTest)
all.equal(credit.pred1, credit.pred2)
all.equal(credit.pred2, credit.pred3)
all.equal(credit.pred1, credit.pred3)
I get the same predictions for all 3 passes. Now, I'm doing this by manually typing the code in the RStudio interpreter. But, If I copy-paste the code from my text editor (which I've posted here: https://gist.github.com/anonymous/32b3c8194362d2e10527), I get an error message saying that there are 3 string differences in the second and third comparison!
How is this possible?

Answer 1

Try using caret's train function:
credit.rf <- train(Class~., data = GermanCreditTrain, method="rf")
instead of
credit.rf <- randomForest(Class~., data = GermanCreditTrain, ntree = 500)
I was able to reproduce the issue and not sure what is causing it. However, the above seems to work when pasted:
credit.rf <- train(Class~., data = GermanCreditTrain, method="rf")
>
> credit.pred1 <- predict(credit.rf, GermanCreditTest)
> credit.pred2 <- predict(credit.rf, GermanCreditTest)
> credit.pred3 <- predict(credit.rf, GermanCreditTest)
>
> all.equal(credit.pred1, credit.pred2)
[1] TRUE
> all.equal(credit.pred2, credit.pred3)
[1] TRUE
> all.equal(credit.pred1, credit.pred3)
[1] TRUE