Filtering Algorithm of Single Table Document Junk Mail Based on the Minimum Risk
Cui Yong-jun · Journal of Lanzhou Jiaotong University · 2009
The limitation of Naive Bayes algorithm used for filtering mails is analysed.The traditional Naive Bayes method doesn't consider the different features between the legitimate mail and the junk in the classification process.On the basis of Naive Bayes algorithm and the Bayes of the minimum risk,a revised E-mail filter is introduced using only spammy words to improve efficiency of the system and reduce the misjudgment of legitimate E-mail,and build a model of E-mail filtering.