overtraining and retraining
Jesus Cea <[email protected]>
| Newsgroups | gmane.mail.spam.spambayes.general |
|---|---|
| Message-ID | <[email protected]> |
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I am a happy Spambayes user for years, and I always see the same pattern: the system works worderfully, and you keep training it with the false positives/negatives. So good so far. After a while the detection rate goes worse, slowly. And training get slower (the probability change after a training cycle moves far slower). My usual approach would to delete the database and retrain again. But this implies a period (a week or two) of mediocre results, even training with "recent" spam. I was thinking about something like going over the database and reduce all counters to 50%, instead of simply starting fresh over. What do you do?. I am using 1.1a4, a bit modded (custom database, etc), should I upgrade to 1.1a6?. I am using SB as a library in my own personal antispam project, not acting as a proxy. PS: My spam threhold is 90%. Maybe i should move it down?. In fact my ranges are ham <20%, 20%<unsure<90%, spam>90%. Are they reasonable?. My "bayescustomize.ini" is: """ [Storage] persistent_use_database:pickle cache_messages:True no_cache_bulk_ham:True [Classifier] x-experimental_ham_spam_imbalance_adjustment:True [Hammie] [Tokenizer] #mine_received_headers:True #train_on_filter:True [pop3proxy] listen_ports:110 remote_servers:127.0.0.1:111 """ - -- Jesus Cea Avion _/_/ _/_/_/ _/_/_/ [email protected] - http://www.jcea.es/ _/_/ _/_/ _/_/ _/_/ _/_/ jabber / xmpp:[email protected] _/_/ _/_/ _/_/_/_/_/ . _/_/ _/_/ _/_/ _/_/ _/_/ "Things are not so easy" _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ "My name is Dump, Core Dump" _/_/_/ _/_/_/ _/_/ _/_/ "El amor es poner tu felicidad en la felicidad de otro" - Leibniz -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQCVAwUBTprzx5lgi5GaxT1NAQLDLgQAhUqUPGh8Y4mKG7CHwNdFc4302oD6PsGm tGESozQ1YWYX1fDglCKhmOXdodNL1ovd3e70or8SGAjDrrtRlXOZhhIMaJbkoFTx SstPlSUm3VmYp8zDw5S0uaaDksl2xErxvwKiRYeC59q4RSUr7cAkfDnL+aWhuabc ZiIGe00csBc= =XQJ6 -----END PGP SIGNATURE----- _______________________________________________ [email protected] http://mail.python.org/mailman/listinfo/spambayes Info/Unsubscribe: http://mail.python.org/mailman/listinfo/spambayes Check the FAQ before asking: http://spambayes.sf.net/faq.html