We use SpamAssassin to classify email and block spam (unsolicited bulk email). SpamAssassin itself uses some tools like:
- Pyzor - Pyzor is a collaborative, networked system to detect and block spam using digests of messages.
- Razor - Vipul's Razor is a distributed, collaborative, spam detection and filtering network
- Bayes Filtering - The Bayesian classifier in Spamassassin tries to identify spam by looking at what are called tokens; words or short character sequences that are commonly found in spam or ham.
Autowhitelisting - The Auto-WhiteList tracks scores from messages previously received and adjusts the message score, either by boosting messages from senders who send ham or penalizing senders who have sent spam previously.
How does it work?
SpamAssassin looks at every incoming e-mail and assigns "SPAM-points" according to many different rules about what SPAM mails contain. For example a mail gets some points if it contains a line written all UPERCASE, if there is a unsubscribe URL or if it speaks about mortgage rates. If the sum of the points surpasses a certain amount, then it is marked as "SPAM" and is rejected