Abstract: This paper applies AIS and Bayes network to the spam filtering, proposes a new spam filtering means based on artificial immune system and Bayes network, and analyses the key problems of the algorithm, and solves these problems and designs affinity’s formula as well as the structure of Ag/Ab. After comparing with AdaBoost algorithm in the PU1 bare corpus, it provides the experimental result finally.

Key words: Spam, Spam filtering, AIS, Affinity, Antigen/antibody(Ag/Ab)

摘要： 将人工免疫系统和贝叶斯网络相结合应用于垃圾邮件的过滤，设计了一个基于AIS和Bayes网络的垃圾邮件过滤算法，分析并解决了该算法实现过程中的关键问题，给出了亲和力的计算方法和抗原/抗体的结构定义。在PU1bare 语料的基础上与Carreras提出的AdaBoost方法进行对比实验，最后给出了实验结果。

关键词: 垃圾邮件, 邮件过滤, 人工免疫系统, 亲和力, 抗原/抗体

CLC Number: 

• huyebowen@163.com