I am working on Bayesian Classifiers. I ve written most of the code myself. Now, I want to test the accuracy of the Classifier. I donot want to train the filter on raw data. Rather I would want to take the tokens and the count of tokens from an established spam filter like SpamAssassin and do it it's way. But the Bayes_toks file of the SpamAssassin is in a hashed form. So is the hammie.db file of the Spambayes. I want the token files in a readable format/plain text. Can anyone plz help me. It's a lil urgent for me.
thanks in advance.