You need a Naive Bayes Classifier. Take a look at PyZor - http://sourceforge.net/apps/trac/pyzor/
Regards, BG On Fri, Dec 23, 2011 at 5:01 PM, Amit Sethi <amit.pureene...@gmail.com> wrote: > Hi All, I am looking for a spam filtering/classification solution . > > The essential requirement are : > 1. Return a negative on all text with certain words. > > But apart from returning negative on certain words. I want it to be able to > classify certain positives. > > For ex - Given the fact that [Sachin, Cricket, Batsmen] are related to > cricket - returns a positive on cricket > Given the fact that [ Inflation, Share Price ] are related to Finance - > returns a positive on finance > > etc ... What kind of tool set/ library can help me achieve that ? > > -- > A-M-I-T S|S > _______________________________________________ > BangPypers mailing list > BangPypers@python.org > http://mail.python.org/mailman/listinfo/bangpypers -- Baishampayan Ghose b.ghose at gmail.com _______________________________________________ BangPypers mailing list BangPypers@python.org http://mail.python.org/mailman/listinfo/bangpypers