Hi,
I'm trying to build a discrete classification system based on a Naive Bayes Learner. All my attributes/features take discrete values. But my test sample instances have values for some attributes which did not occur in my training sample. Please let me know if there is a way to handle occurrence of new values in test sample.
Is laplace smoothing included in the naive bayes implementation.??? This would help.

Regards,
Praveen Karri.