Paul Houlbrooke wrote: > Paul Houlbrooke wrote: >> When I use the analyser through the email interface, it sends the email >> with all the ASSP headers and the new ASSP subject (i.e. [Bayesian] >> Subject]). This produces a probability of 0 and hardly any bad words. >> When I run the analyser from the web GUI and remove the ASSP headers, I >> get different results (i.e. more bad words) but still a probability of >> 0. Only after I remove the ASSP headers and [Bayesian] from the subject >> line through the web GUI do I see a correct analysis with a probability >> of .2020. So doesn't this make the email interface for the Mail >> Analyser, for lack of a better word, pointless? The only way I can get a >> correct analysis is to manually remove the ASSP stuff from the email and >> submit it via the web GUI. >> >> 1.3.5 (19.14) >> > > Oh, and in the log for the same email the probability was reported as 1.00. > > Apr-2-08 07:55:14 id-37313-02526 70.62.109.194<[EMAIL PROTECTED]> to: > [EMAIL PROTECTED] PB-Message-Score is 85, added 45 (Bayesian > Probability: 1.0000) > >
Actually, the griplist had a value of .83, which when added to the .2 would be 1.00 (ish). So that may explain that part at least. • 10.0.0 has a Griplist value of 0.838689: (adds 0.838689 0.838689) ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ Assp-test mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/assp-test
