Hi Markus, when using standard data mining methods from R, we usually recommend using the R Learner and R Predictor nodes. For an example, connect to the EXAMPLE server in the KNIME Explorer and download the 006_R Integration/006_001_R_example workflow into your LOCAL repository. Important to know is that the R Learner works on a data frame called variable R, that is, your dataset is contained inside R, and produces its result again in a variable R; for example R<-rpart(R$target~., R) would generate a decision tree on R using variable target and returns it in a variable R. The R Predictor makes this model available inside RMODEL and the test data inside the variable RDATA. Hope this helps.