POPFile

POPFile is free e-mail classification tool. The underlying classification technology is based upon a Naive Bayes classifier that uses feedback to "learn" your personal e-mail classifications. POPFile learns from example and can then extrapolate based upon your feedback to correctly classify new e-mail messages. It is ideal for classifying spam with a very high degree of success after only two or three days of training. If you have e-mail, you probably receive spam, and POPFile is a very effective spam antidote.

Why does it work?

The classifier within POPFile doesn't use human-composed rules to classify e-mail. Instead, it "reads" your incoming e-mail, assessing the language within it in order to classify it appropriately. The beauty of this approach is that POPFile can learn to recognise the "hard sell" language often found in spam e-mail. Once POPFile has learnt to recognise selling language it can identify spam as soon as it arrives without requiring any manual intervention from anyone, least of all you. Spammers can still get past POPFile by using language that doesn't talk about selling things but, of course, selling is what spam is all about. As a result, POPFile can be very successful in protecting you from unwanted e-mail.

We use it ourselves

We receive hundreds of spam e-mails each week, and POPfile accurately identifies around 99% of spam, which means we waste almost no time managing the e-mail we don't want to see.

How to configure POPFile to beat spam

POPFile comes with reasonably good instructions on how to install and configure it but we thought that it would be worthwhile providing the 10 minute setup routine to get you started. Once POPFile is in and working you can play around with the additional capabilities and really optimise your e-mail management.

1. Download and install POPFile

Download POPFile and run the installation program. If you are using Microsoft Outlook or Microsoft Outlook Express, POPFile can configure everything automatically during the installation process.

2. Create some "Buckets"

POPFile classifies incoming e-mail into "Buckets". Initially, we only want two buckets: spam and notspam. To create these, start up the POPFile user interface and select the "Buckets" tab. Near the bottom of the page, you can find the tools to create your buckets.

3. Turn off "Subject Modification"

By default, POPFile modifies the subject of every incoming e-mail to include information about the bucket it has been classified into. We don't want the notspam e-mails to be modified so turn off Subject Modification for the notspam bucket. You can change the colours here if you like, too. We use red for spam and green for notspam but you can use whatever colours you like.

4. Start reading e-mail

POPFile is now ready to go. When you start reading e-mail with your e-mail program, POPFile intercepts everything before it gets to your Inbox. Anything it thinks is spam gets put into the spam bucket and the e-mail comes through with the subject prefixed with [spam]. Initially, most of the classifications will be wrong because POPFile hasn't been trained.

5. Train POPFile

Whenever POPFile incorrectly classifies an e-mail message, you need to tell it the correct classification. Do this through the POPFile user interface History tab. Find the incorrectly classified message, select the bucket it should have been classified into, and Reclassify. After it has intercepted about 100 messages, POPFile will become pretty good at classifying your e-mail correctly.

6. Finally, use rules to ditch your spam

Use the Rules capability in your e-mail client (e.g. Outlook Express Rules Wizard) to identify spam messages based upon the subject (ie: "subject" contains [spam]) and either delete them (if you're brave) or move them into a spam folder for periodic review. Once the rule is activated, you will no longer see spam in your Inbox!