@cupcake wrote:
Hi,
I have faced this block for a few days even with much research online, and much grateful if there are some insights in dealing this kind of warnings: " Contrasts dropped from factor x" . This occurred when I tried to use newly created decision tree model to predict on the test data using rpart package.Following are the scripts that I have used :
db$staffx.f<- as.factor(db$staffx)contrasts(db$staffx.f) = contr.treatment(8) .... create same for other categorical variables.
tree<-rpart(L2 ~ staffx+....other variables, method = "class", data = db )
printcp(tree)
rpart.plot(tree,type = 3,tweak =1.2)
cpok<-tree$cptable[which.min(tree$cptable[,"xerror"]),"CP"]
tree.pruned<-prune(tree,cp=bestcp)
pred1 <- predict(tree,Testdata,type="class")
In which areas that I need to adjust ? Does this mean the model has failed?
Many thanks!
Posts: 1
Participants: 1