Ubuntu - SpamAssassin - Train SpamAssassin with Ham and Spam

Train with Spam.

sa-learn --showdots --mbox --spam spamfolder

Train with Ham.

sa-learn --showdots --mbox --ham hamfolder

NOTE: SpamAssassin needs at least 200 Ham and Spam message before it becomes fully trained. Everything above just makes it “safer” as in “more valid” or “accurate”. With auto-learning these numbers will increase over time and also decrease as statistics of old mails will be dropped over time.

Automate learning

This learning can be automated by:

Create a dedicated email account for Spam and another dedicated account for Ham.
When a user receives any Spam they should forward this onto the inbox of the Spam email account.
Users can also copy non-spam email to the inbox of the Ham email account.
Run a cron job to train against the Spam in the Spam email account.

The script for the Spam can be something like:

/usr/bin/sa-learn -u <username> --spam /home/domain/<user>/Maildir/.<folder name>/cur/  
/usr/bin/find /home/domain/homes/<user>/Maildir/.<folder name>/cur/ -mtime+28 -exec rm {} \;

WARNING: Do not have the <user> and <folder name> point to account of an actual user.

This should instead point to the the email account that holds the spam.

Have ham learning be done as:

/usr/bin/sa-learn -u <username> --ham /home/domain/<user>/Maildir/.<folder name>/cur/

WARNING: Do not have the <user> and <folder name> point to account of an actual user.

This should instead point to the the email account that holds the Ham.