Spam Detection Approach Based on C-Support Vector Machine and Kernel Principal-Component Analysis

Current spam detection algorithms have poor generalization ability as given small samples and less priority knowledge. This paper proposed a spam filtering detection protocol based on Kernel principal-component analysis (KPCA) and C-Support Vector Machines (C-SVM) which can solve and implement the mentioned problem. Compared with the traditional algorithms this method can achieve higher detection rate and improve detection efficiency, and be easily generalized in practice. At last the experiment on data set shows the effectiveness and excellent performance of this method.