Bayesian Filtering Algorithm

Implemented algorithm for spam filtering is based on Bayesian approach. Using the algorithm Spam Reader determines the score corresponding to the probability for a particular message to be spam. This score is used to decide whether a message is spam or not. This approach to spam recognition is one of the most natural and thus precise and reliable. Regular updates and self-training option, which make filtering algorithm very flexible, guarantee your mailbox a really reliable anti-spam protection.

To determine the score for a message Bayesian Filtering Algorithm uses the special spam database. When you install Spam Reader on your computer it already has the default database of spam and non-spam words.

To make spam recognition highly effective for your particular mailbox Spam Reader analyzes your personal correspondence to update the spam database. On the first run of Spam Reader you will see Bayesian Filter Update wizard. It suggests you to allow Spam Reader to analyze your correspondence in order to update the spam database and thus increase the accuracy of spam recognition for your personal mailbox. After that Spam Reader will constantly update the base when you work your correspondence. When you click "Good E-mail" or "Junk E-mail" button, the selected messages are automatically analyzed to update the base.