I have a question regarding NA in randomForest (in R). I have a dataset which include both numerical and non-numerical variables, and the data includes some NA. Do anyone have some tips how to deal with this? I tried to use na.roughfix but then i get an error message "na.roughfix only works for numeric or factor". I also tried rfImpute but this does not work either because I have some NA in my response variable. Does anyone have som tips?