What criteria does MailDefender use to detect spam?
MailDefender defines spam using two main methods: Blocking e-mail from known spam senders and Bayesian statistical analysis. Firstly, it checks a set of block lists that contain the IP addresses of known spammers’ outfits, open mail relays and blocks of dynamic IP ranges that have previously been used as spam hosts. Secondly, it analyses the content of the e-mail against a training set of known good e-mail and known spam e-mail. It does this by analysing the mail’s wording in context, and comparing this with the same wording in the same context in the two training sets. Using this analysis MailDefender scores the e-mail as being either good e-mail or spam.
