Using LPP and LS-SVM for spam filtering
Xia Sun, Qingzhou Zhang, Ziqiang Wang · 2009
To efficiently deal with spam mail filtering problem, a novel spam filtering algorithm based on locality pursuit projection (LPP) and least square version of SVM(LS-SVM) is proposed in this paper. The mail message features are first extracted by the LPP algorithm, then the LS-SVM classifier is used to classify mails into spam and legitimate. Experimental results demonstrate that the proposed algorithm performs much better than other related spam filtering algorithms.