What is Bayesian spam filtering?

Bayesian filtering applies probability theory to classify messages. It learns which words appear frequently in spam versus legitimate email, then calculates the probability that new messages are spam based on word frequency.

Training: the filter analyzes known spam and legitimate messages, building probability tables for each word. Classification: new messages are scored based on cumulative word probabilities.

Strengths: adapts to individual user patterns, learns new spam vocabulary automatically, and handles misspellings well. Weaknesses: vulnerable to poisoning (attackers diluting word probabilities), requires training data, and can't evaluate non-textual signals.