data("iris")  ;  DT <- iris  
nr <- nrow(DT)
sa <- sample(1:nr,nr) ; DT <- DT[sa,]


ma <- matrix(data = rnorm(1500) , ncol = 10) # make noise
clnm <- rep(NA,ncol(ma))
for(i in 1:ncol(ma))     clnm[i] <-  paste0("noise",i) 
colnames(ma) <- clnm


DT <- cbind(ma,DT)  ; rm(ma)
#\\\\\\\\\\\\\\
Species <- DT[,ncol(DT)]
DT <- DT[,-ncol(DT)]
train <- 1:100 
test  <- 101:150 
#\\\\\\\\\\\\\\
library(caret)
normal <- preProcess(DT,method = c("center" , "scale"))
tr <- predict(normal,DT[train,])
ts <- predict(normal,DT[test,])
#\\\\\\\\\\\\\\
library(randomForest)
RF <- randomForest(Species[train]~., tr, importance=T) 
pr <- predict(RF,ts)
#\\\\\\\\\\\\\\
library(ForeCA)
forec.dt <- as.matrix(  rbind(tr,ts)  )
fo <- foreca(forec.dt, n.comp = 4)

F.RF <- randomForest(Species[train]~., fo$scores[train,])
f.pr <- predict(F.RF,fo$scores[test,])



confusionMatrix(Species[test], pr)  # simple iris
confusionMatrix(Species[test], f.pr) # foreca