Generating Behavior-based Classification Rules for Spam Filtering Using Enhanced Induction Trees

Abstract

We present in this paper a novel featuring method for rule-based spam filtering. Instead of classifying emails according to keywords, this study analyzes the spamming behaviors and extracts the representative ones as features for describing the characteristics of emails. An enhanced decision tree algorithm with weighted information gain is proposed, which builds decision trees by considering the importance of behavior-based features revealed from emails. Since spamming behaviors are infrequently changed, compared with the changing frequency of keywords used in spams, behavior-based features are more robust with respect to the change of time; so that the behavior-based filtering rules outperforms keyword-based filtering ones. The experimental results indicate that our method is more useful in distinguishing spam emails than that of keyword-based comparison.