An analysis technique that has been applied to eliminating spam. It "learns" to differentiate real mail from advertising by examining the words and punctuation in large samples of both types of messages. Based on the probability theorem of Reverend Thomas Bayes (1702-1761), it selects a set of words and numbers ("tokens") from the text and compares their ratio between good mail and spam. Using the tokens, the Bayesian approach looks at new mail and calculates the probability that the message is bogus. See
spam filter.