Yes, the "Values" box will accept multiple entries. Put one entry per line.

You should use magnets with care. If you are just starting to use POPFile you may think that magnets are the way to go. But if you set up a magnet for everyone you communicate with, all those messages will bypass the Bayes engine and POPFile will never learn what a good message looks like. If you train POPFile, however, it will learn what words are used in the messages from your friends and if one of them gets a new email address (or if you win a new friend), POPFile will have no trouble at all correctly categorizing this piece of mail that your magnets never would have caught.

There are occasions where a magnet might actually help you, however. Suppose you are on a mailing list where spam is discussed. A message to that list might contain a complete spam mail and POPFile might thus have a hard time to tell whether all those spammy words actually make this a spam message or not.

Use magnets with caution as POPFile cannot learn from any of the email caught by a magnet. Years of experience with POPFile has created a consensus that magnets should be used carefully and infrequently. While additional buckets do not generally affect POPFile's accuracy, additional magnets usually have a deleterious effect on POPFile's accuracy.

I have been using POPFile for over six years. In that time I have never used magnets to help classify my email into 7 buckets ('spam' plus 6 others for good email). My current accuracy is 99.62% (42 errors out of 11,234 messages).

Brian

See also:
What is a magnet?
What variables affect 'good' statistical ratings?
Does POPFile support Whitelists? Can I import an address book into POPFile?
TOE (Train Only on Errors)