I cannot process large datasets - any ideas?

Since most schemes in WEKA need to have all of the data present in memory, large datasets can be a problem. The article Classifying large datasets tries to present some solutions. But make sure that you have already read how to deal with an OutOfMemoryException.