Is there a way to shrink the size of RandomForest-class (an S4 object), so that it requires less memory during run-time and less disk space for serialization? On my system the data slot is about 2GB, which is causing problems, and I'd like to see whether predict() works without it. # example with a much smaller data set (i.e., less than 2GB) require(party) data(iris) cf <- cforest(Species ~ ., data=iris) str(cf, max.level=2) cf at data <- NULL # this fails Andrew