New script to train bogofilter
Boris 'pi' Piwinger
3.14 at logic.univie.ac.at
Mon Jun 30 17:25:01 CEST 2003
Hi!
I wrote a perl script which trains bogofilter on error. It
produces very small databases. We'll have to see how good
that works. Any comments are warmly welcome.
Here is a sample run:
[3.14 at pi ~/local/bogolists]$ build-bogofilter-database.pl \
.bogofilter 'ham*' 'spam*'
Starting with this database:
(empty)
Done:
Read 22457 ham mails and 14806 spam mails.
Added 196 ham mails and 202 spam mails to the database.
spam good
.MSG_COUNT 202 196
False negatives: 232
False positives: 52
[3.14 at pi ~/local/bogolists]$ build-bogofilter-database.pl \
.bogofilter 'ham*' 'spam*'
Starting with this database:
spam good
.MSG_COUNT 202 196
Done:
Read 22457 ham mails and 14806 spam mails.
Added 69 ham mails and 91 spam mails to the database.
spam good
.MSG_COUNT 293 265
False negatives: 68
False positives: 20
[3.14 at pi ~/local/bogolists]$ build-bogofilter-database.pl \
.bogofilter 'ham*' 'spam*'
Starting with this database:
spam good
.MSG_COUNT 293 265
Done:
Read 22457 ham mails and 14806 spam mails.
Added 17 ham mails and 48 spam mails to the database.
spam good
.MSG_COUNT 341 282
False negatives: 1
False positives: 6
pi
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: build-bogofilter-database.pl
URL: <http://www.bogofilter.org/pipermail/bogofilter/attachments/20030630/4ddb4da0/attachment.pl>
More information about the Bogofilter
mailing list