Message Classes

In SpamExperts we use different message classifications to describe why a message was treated the way it was by the filter.

You can check message classifications from inside SpamExperts by clicking Classifications in the top section of the Incoming or Outgoing Log Search pages, which will provide you with a panel detailing what classifications mean.

Example

The main class that displays for a piece of mail defines what has been recognized in the message.

The sub class is why that main class has been attributed to the message.

For Example:

These messages have the following class details:

  • Main Class - not-spam
  • Sub Class - statistical-method1-cluster & statistical-method1-global
  • Extra Class - 0.03833113230835744 & 0.00004507188299779363
  • Status - Delivered

These messages are delivered to the recipient with a main class of Not-Spam because they have the sub class of statistical-method1-cluster and statistical-method1-global (which means they are statistically similar to messages received either to the cluster or globally before that have been trained as not-spam) and so their combined score - in the Extra Class column - is below the Beneficial to train threshold.

Main Classes and Sub Classes

Error Class

Messages that have been trained as spam or not spam will show an Error Class of either:

A message that has been reported to the filtering system as a false positive (i.e. Trained as Not-Spam).

This message which was caught as spam when it should not have been.

A message that has been reported to the filtering system as a false negative (i.e. Trained as Spam).

This message which was received as legitimate when it should not have been.