改进的人工免疫垃圾邮件过滤算法

Improved artificial immune algorithm applied for spam filtering

针对AISEC模型中自体库和基因库生成效率不高的弊端,提出基于词频和MI互信息的自体库和基因库生成算法,同时对自体库和基因库的更新策略进行改进。实验结果表明,应用改进后的算法至少可以将邮件分类时间缩短10%,同时在虚报率方面得到了明显改善。

AISEC model is not so effective in generation of self library and gene library.It improves the efficiency of self library and gene library generation by adding MI and the frequency of words,and improves the auto-update strategy of self library and gene library.Experiments show that the improved algorithm can reduce at least 10% of the time,and can reduce the fallout rate.