论文部分内容阅读
受生物免疫系统工作机制的启发,本文提出一种基于免疫原理的个性化 Spam 过滤算法.其主要思想是根据用户兴趣和邮件特征定义垃圾邮件社区,将各垃圾邮件归类于不同的垃圾邮件社区,抽取各个垃圾邮件社区的特征并用一组特征检测器来表示,检测时通过判断待检测邮件是否归属于某垃圾邮件社区来进行过滤.该算法是一个增量学习算法,能连续过滤垃圾邮件.算法中免疫学习与免疫记忆机制的采用不仅能提高垃圾邮件过滤的检出率与正确率还能加快邮件过滤的速度.文中通过测试实验和分析表明,本文算法的垃圾邮件过滤性能优于AISEC 与 Naive Bayesian 算法.
Inspired by the working mechanism of biological immune system, this paper proposes a personalized Spam filtering algorithm based on immune principle.The main idea is to define the spam community according to the user’s interest and the characteristics of the email, and classify each spam into different spam communities , Extracts the characteristics of each spam community and uses a set of feature detectors to represent it, which is judged by judging whether the message to be detected belongs to a spam community. This algorithm is an incremental learning algorithm that can continuously filter spam. The adoption of immune learning and immune memory in the algorithm can not only improve the detection rate and accuracy of spam filtering but also accelerate the speed of mail filtering.Through the test and analysis, it is shown that the spam filtering performance of this algorithm is better than AISEC and Naive Bayesian algorithm.