E-mail classification by SVM optimized with genetic algorithm
Wei Wang · Journal of Computer Applications · 2009
A method of classifying E-mail by Support Vector Machines(SVM) and Genetic Algorithm(GA) was proposed.In the first step,the mails were preprocessed,and then the combination of support vector machine parameters of C and kernel function parameters was optimized by genetic algorithm.Finally,the E-mail was classified by the optimized SVM.The experiments on a data set composed of 5 800 mails show that the precision is 89.67%,which indicates that this method indeed improves the accuracy of filtering Chinese spam.