This is a method of using many decision trees. There is a package named "random forest" in R. It is unclear if you should consider only the first 10 numeric attributes or not. You may want to try both. If the data set is too large to be handled, try some subsets only and gradually increase the size.