An Intelligent Spam Email Filtering Approach Using a Learning Classifier System.

Emails have become a common modality to exchange messages and information over the internet. Some emails are frequently received from malicious senders, which can lead to several problems. Thus, there is an urgent need to develop reliable and powerful methods for filtering such emails. In this paper, we present a new approach to filter emails to spam and ham based on the text of the email being analyzed. After feature extraction, a supervised classifier system is adopted to generate a rule-based model that captures spam and non-spam patterns used later as an interactive filter. The aim is to obtain a compact model that provides a highly accurate performance using only a few rules. The proposed approach was applied to two datasets in terms of size and features. Experimental results indicated that an accuracy of 99.7% was achieved using a model with only 74 rules and 70 conditions. With this reduced number of rules, the proposed approach presents an efficient solution for filtering spam email using the trained model. In addition, the results of the proposed approach showed better performance than their counterparts.