OSBF-Lua Stuff

OSBF-Lua is a Lua module for text classification written by Fidelis Assis which includes an extremely fast and effective spam filter.


train_osbf trains the OSBF-Lua spam filter using the e-mails from a given ham and/or a given spam folder (both of them are expected to be in mbox format). Note that OSBF-Lua learns so quickly that a pre-training of the spam filter won't be necessary in most cases. However, train_osbf might be useful for ISPs, for test purposes or for other special cases. It could also be used for training on error and reinforcements during regular operation of OSBF-Lua, but see Christian Siefkes' trainspamfilter as an alternative solution for this purpose. See the output of train_osbf.lua --help for the available command line options.

Download: train_osbf.lua


resize_databases is a trivial script which resizes the OSBF-Lua spam filter database files "spam.cfc" and "nonspam.cfc". The new number of buckets can be specified on the command line (use a prime number!). The default size is 500501 buckets.

Download: resize_databases.lua


OSBF4Mutt provides macros for training the OSBF-Lua spam filter from within Mutt. Also, messages with a score within OSBF-Lua's reinforcement zone are marked specially in Mutt's folder index until they have been trained and Mutt's spam attribute is defined. See the README file for details.

Download: osbf4mutt-0.2.tar.gz

Valid XHTML 1.1!  Valid CSS!  Powered by Debian GNU/Linux

Last modified: $Date: 2013/01/28 21:32:16 $ by Holger Weiß