垃圾邮件过滤系统的探究与实现

Research and implementation of spam filtering system

电子邮件已成为现代通信中不可缺少的一部分,但垃圾邮件的日益泛滥给计算机系统安全和人们的工作与生活带来了极大的威胁,反垃圾邮件已成为一个非常重要的任务。在传统的黑白名单过滤技术的基础上,引入了IP信誉评分机制,并结合基于规则的过滤技术和基于内容的贝叶斯过滤技术,从而建立了一个多层次的垃圾邮件过滤系统模型。同时在系统中应用了反馈学习技术,以弥补因误判而造成的损失和提高系统的准确率。经实践验证,本系统适用于用户终端使用,有较高的可行性。

E-mail has become indispensability of modem communications, but a large ofspam flood has seriously threated the computer system security and people＇ s work and life, so anti-spam already become a very important task. A sub-level of spam filtering system model based on the traditional black and white list filtering technology is established. The model also introduce the IP credibility score mechanism, combined with rules-based filtering technology and content-based Bayesian filtering technology. At the same time, feedback learning techniques is applied to compensate for the losses caused by misjudgment and improve the system accuracy. The practice verified the system is suitable for end users in a higher degree of feasibility.