Research of Mail Filtering Based on Bayesian Incremental Classification
Yu Chengyi · Science Technology and Engineering · 2009
With the rapid development of internet,spam is widely spreaded.Facing the increasingly serious situation of junk mails.The Bayes mail filtration model is proposed,and the application of Bayesian classification method based on the filtration of junk mail is discussed.In view of difficulty to acquire the category label in great quantities of training regulations problem and Bayes feature of incremental study.An incremental bayesian classification is presented based on small training set.Through minimizing the classified loss of current sorters,the samples which can enhance the performance of sorters is added to the training regulations.The experiment results show that the method is feasible and effective.