Adaptive Message Filtering Spectrum

Currently, we use adaptive message filtering to filter out spam. The adaptive filter usually applies a percentage value to incoming messages, and if a message’s percentage value goes below a threshold, the message is considered spam.

But what if we could use these percentage values to do more than just mark messages as spam or not? Certain adaptive filters already provide the ability to view a message’s percentage value, there are great opportunities if we could take advantage of it.

For example, you could sort your unread messages by greatest percentage value, thus getting what arguably are your most important messages first. You could also adjust your spam threshold at any time to accommodate message overload. Finally, you can tell when an important kind of message is being interpreted by your filter at a very low threshold, enabling you to correct it before it gets any worse.

By providing the client with the ability to view and use the percentage value by an adaptive message filtering system, we can offer the user greater control over their incoming messages.

Update: It looks like FogBuzz 4.0 will have something like this. They apparently expand it to do multiple facets, rather than just a spectrum like shown here.