ABSTRACT
This chapter presents a novel feature construction approach based on term space partition (TSP), which aims to establish a mechanism to make terms play more sufficient and rational roles in e-mail categorization [132]. First, motivation of proposing the TSP approach is described. The main principle of the TSP approach is then introduced. Detailed implementation of the TSP approach for spam filtering is given next, including preprocessing, term space partition, and feature construction as three core steps. Finally, conducted experiments are shown to indicate the effectiveness of the TSP approach.
