How URL Spam Filtering Beats Bayesian/Heuristics Hands Down

Overview

The evolution and sophistication of spam and spammers has sparked much debate as to which spam filtering methods are the most accurate. This whitepaper addresses issues and problems associated with Bayesian and Heuristic filtering solutions and argues that URL ("click me" link) filtering is the most accurate and predictable method of blocking spam. URL filtering does not require customer tuning and eliminates the need for Bayesian/Heuristic solutions which are error prone and unpredictable. Implementing URL filtering with conservative IP blacklists and very limited and specific heuristic filtering has the ability to solve spam with over 99% blocking, miniscule false-positive and without any customer tuning.