Email filtering based on structural feature analysis and text classification
Meijuan Yin · Jisuanji gongcheng yu sheji · 2010
In many specific applications such as E-mail mining,social security management and so on,a large number of irrelevant E-mail data seriously interferes with the effects of the target processing task.For this problem,an integrated method is designed combined with traditional spam filtering technology.It initially filters E-mails according to some irrelevant structural features.Considering the different contributions of subject and body text to the category of an email,the algorithm of naive Bayesian E-mail classification is improved to filter again.The experimental results show that the method makes up the disadvantage of spending too much in text classification and improves the precision and fallout which verifies the practicality and feasibility of the method.