On Tue, Mar 29, 2011 at 10:40:30PM +0200, Tom Hendrikx wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On 29/03/11 20:44, Kenneth Marshall wrote: > > On Tue, Mar 29, 2011 at 09:40:19PM +0300, Ibrahim Harrani wrote: > >> Hi Ken, > >> > >> I added the list to dspam.conf. > >> Should/Can I delete from database those pattern as well? > >> > >> Thanks > >> > > > > Unfortunately, only the hash of the header is stored so there > > is not a way to identify the applicable tokens. You may want to > > regenerate the corpus. If you create the table with a <100% > > fillfactor, the training should go faster. > > > > That is incorrect. You can easily recreate the hash with dspam_crc, f > you know the original token: > > $ dspam_crc "Received*from+userid" > TOKEN: 'Received*from+userid' CRC: 2977181303331328604 > > vmail=# select * from dspam_token_data where token = '2977181303331328604'; > uid | token | spam_hits | innocent_hits | last_hit > - -----+---------------------+-----------+---------------+------------ > 1 | 2977181303331328604 | 4 | 2203 | 2011-03-29 > > > - -- > Regards, > Tom
Yes, that is true. Unfortunately, you have to know all of the original tokens for run dspam_crc. I guess you can grab them from the original training data and delete them from the database using dspam_crc. Cheers, Ken ------------------------------------------------------------------------------ Enable your software for Intel(R) Active Management Technology to meet the growing manageability and security demands of your customers. Businesses are taking advantage of Intel(R) vPro (TM) technology - will your software be a part of the solution? Download the Intel(R) Manageability Checker today! http://p.sf.net/sfu/intel-dev2devmar _______________________________________________ Dspam-user mailing list Dspam-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dspam-user
