Summary

The following is a summary of the Bayesian Filtering Subsystem.

Database Maintenance

Bayesian database has to be kept updated by teaching it spam and ham in fairly equal quantity.

Seeding the Database

The Bayesian Filter will not run until it is taught with a minimum of 200 spam and 200 ham messages.

Learning Email

Upload ham and spam in mbox format as described on page , to train the Bayesian database.

Forgetting Previously Learned Email

The Bayesian Filter can forget incorrectly learned spam/ham. See page for information about forgetting mail.

Auto Learning

The Bayesian Filter can automatically learn from ham and spam passing through the spam filter, after the Bayesian database has been seeded. To enable this option, see page .

Rebuilding the Database

You can rebuild the Bayesian database as described on page .

Forcing an Expiry Run

You can expire old information in the Bayesian database, described on page .