HADES™ Outlook Spam Filter: An Anti-Spam Software

Abstract

Hades a Spam filtering software is a research oriented commercial project that focuses on discovery and implementation of an accurate technique for filtering Spam. The aim of this project is to get a large repository of Spam emails and find most important attributes of a Spam email and come up with a statistical model which will later be used for Spam detection. The research is focused on finding a technique that can cater for dynamically changing practices employed by Spammers to breach existing Spam filters. Every time a Spam filtering software is developed spammers discover ways to bypass the filter, the only thing that can best decide whether the mail is Spam or not is human intelligence that’s why we considered Artificial Intelligence and neural networks to solve this problem. Finally we find a neural network technique Bayesian networks to develop an accurate and dynamic filter for Spam. The main advantage of using Bayesian is self-learning and accuracy. Hades Anti Spam filter consists of three major modules:- • Training Phase • Analysis Phase • Learning Phase During training phase Hades is trained the most important attributes of a Spam mail. During Analysis Hades Bayesian filter uses its learned knowledge to decide whether the email being analyzed is Spam or not. During learning phase Bayesian filter updates its knowledge base with the new attributes and these new learnt attributes are used to make future analysis.