I ever train a naive bayes classifier by a large training size. like dmoz , using lingpipe package. but out of memory. i.e., exceed limit of java heap size.
I want to know does any one tried a big training size to train a mahout bayes classifier for text ? thanks. -- Li Jun
