Does POPFile support Whitelists? Can I import an address book into POPFile?

The short answer is you can not import addressbooks into POPFile. The good news is that you probably do not need to use whitelists because using POPFile doesn't preclude you from using whitelists in your email client–if you want to. That is, POPFile classification can coexist quite well with client-based rules. For example, both Outlook and Eudora allow rules that filtering if the “From:” field intersects someone in your address book.

In fact, mail-client based whitelists combined with POPFile classification can be a powerful combination. Allowing your email client to maintain your contact list, you eliminate the duplication of effort in maintaining a separate POPFile whitelist. See your email client's documentation for information on creating filters (sometimes called “rules”).

Detailed Explanation

This question comes up from time to time–usually when users set up POPFile for the first time. The reason POPFile doesn't include a way to import an address list is that experience shows that whitelists alone are rarely a very effective way of dealing with email classification or spam. The whole idea of a Bayesian filter is that once trained, the sum of the probability of all of words in an email (and the header) will classify the email correctly usually >98-99% of the time. POPFile will learn from your classifications and generally not become more accurate with the “help” of whitelists.

As mentioned, you can use some mail clients to create a whitelist if you think it is necessary–but after 800-1000 emails POPFile will reach very high efficiency. See POPFile accuracy.

There are some good reasons why whitelists may not be a good idea. Here are a couple cases that show problems with them:

1. The first email from each contact inevitably means they aren't on your whitelist–that's a potential problem. Note a Bayesian filter like POPFile, can correctly classify messages even when they come from a sender it has never seen before because it is using a statistical analysis of the occurrence of all words in the email (and header).

2. Many computer worms send mail from people on your whitelist that may actually be spam. POPFile's classification is superior in this scenario too since it looks not just at who sent the email, but the content as well.

There is no getting around the fact that you have to check POPFile's classifications–particularly during the initial learning phase–and periodically after that. However, you should see excellent accuracy in short order.

Note: POPFile includes ” magnets” which can be used to allow email matching a narrow criteria to bypass POPFile[1]. As the POPFile documentation's Magnet page suggests, use magnets with caution as POPFile can lean from none 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.

[1] Magnets are limited to matches to the “TO”, “CC”, “FROM” and “SUBJECT” fields.

 
faq/whitelists.txt · Last modified: 2008/02/08 19:49 (external edit)

Should you find anything in the documentation that is incomplete, unclear, outdated or just plain wrong, please let us know and leave a note in the Documentation Forum.

Recent changes RSS feed Donate Driven by DokuWiki
The content of this wiki is protected by the GNU Fee Documentation License