【摘要】 随着电子邮件的广泛使用,垃圾邮件的危害日益增大,用户的个性化需求也日趋强烈。文中提出了一个基于Agent的邮件过滤与个性化分类系统,既能过滤垃圾邮件,又可以根据用户的个性化需求对正常邮件进行自动分类。垃圾邮件过滤采用了朴素贝叶斯方法,邮件的个性分类采用了最小风险贝叶斯方法。邮件个性化分类有效地利用了邮件过滤的输出,提高了系统运行的效率。本系统还可以接收用户的反馈并传递给对应的Agent,从而改进分类算法,不断地微调分类系统。更多还原

【Abstract】 With the wide application of E-mail,the spam does more and more hurt, and the users want more personalized services.A mail filter and personalized classification system based on Agent is provided.This system can not only filter the spam,but also classify the normal mail automatically according to the users’ personalized requirements.The spam filter uses naive Bayes method,and the mail personalized classification uses minimum risk Bayes method.The mail personalized classification makes good use of the output of the spam filter,as a result improves the system’s operating efficiency.This system can get users’ feedback and send to the corresponding agent,in order to improve the classification algorithm and make delicate adjustment to the classification system.更多还原