【作者】 谢嵘；

【导师】 吴伟民；

【作者基本信息】 广东工业大学 ， 计算机应用， 2005， 硕士

【摘要】 随着互联网的普及,电子邮件正成为一种最快捷、最经济的通信手段。而垃圾邮件占用大量传输、存储和计算资源,影响了网络的正常运行;它严重干扰了人们正常生活,甚至造成很多额外的经济支出和安全隐患。据统计,全世界的公司企业每年大概要花费80亿至100亿美元来解决垃圾邮件问题,并且这一数字每五个月会翻一番。因此,对反垃圾邮件技术的研究已经成为影响互联网发展的重要课题之一。 目前对付垃圾邮件的主要方法是基于规则的过滤,这种方法存在查全率低、容易出现误判、适应性差、维护麻烦、无法阻挡病毒通过邮件传播、不支持中文邮件等不足。针对目前垃圾邮件过滤系统的不足,如何提高垃圾邮件过滤系统的查全率和降低误判率,如何改善邮件过滤系统的适应性,以及对邮件进行病毒检测是本课题需要解决的问题。这些问题的解决是实现本文研究目标的关键。 本课题以垃圾邮件过滤为目标进行研究,设计并实现了一个垃圾邮件过滤系统SpamBlocker。该系统整合规则过滤、贝叶斯分类、病毒检测和黑/白名单等垃圾邮件检测技术,采用评分方法判断邮件的垃圾性,并利用规则过滤给贝叶斯分类提供学习样本,提高了系统对新垃圾邮件的适应性。通过测试,SpamBlocker邮件过滤系统具有很好的反垃圾邮件能力,并被成功应用到实际运行的系统中。 本文的内容组织如下:第一章介绍了垃圾邮件的概况以及反垃圾邮件技术,并介绍了本课题主要研究工作;第二章介绍了SpamBlocker系统使用的几个相关技术;第三章给出了SpamBlocker系统的的总体设计,第四章和第五章分别介绍了MailFilter服务器和MailAdmin客户端的设计与实现,着重介绍了SMTP协议、病毒邮件检测、规则过滤器、贝叶斯分类和日志分析等模块的原理、设计与实现。最后一章给出了SpamBlocker系统测试方案及测试结果。更多还原

【Abstract】 With the popularization of Internet, E-mail becomes one of the most convenient and economic communication methods. However, due to occupying large amount of transferring, storing and computing resources, spam not only greatly influences the common life of people, but causes many economic losses and security problems. According to the statistics, spam costs about 8$ to 10$ billion for the company all over the world to eliminate spam annually, and the cost will double every five month. So, the research of anti-spam technologies becomes one of the important tasks which will greatly affect the development of Internet.Currently the main method handling with spam is rules-based filtering. The disadvantages of this method are low accurate in checking and prone to inaccurancy, bad adaptation, bothering maintenance, incapability of blocking mails with virus, be short of the support for Chinese mail, etc. Aiming at the deficiency of the filter system against spam, how to increase the accuracy of filter system against spam how to decrease the inaccuracy against normal mails, how to improve the adaptation of filter system against spam, and the methods for scanning virus in mails are our main concerns in this thesis. And the solution of these problems is the key point in our research approach,These subjects are concentrated on the filtering against spam, further, design and implement a filter system against spam named SpamBlocker. The product integrates the inspection technology for filtering spam such as rules-filtering, Bayes classification, virus scanning and black/white list, etc. Furthermore, it adopts the rules of scoring to determine which mail is spam and the filtering rules to provide Bayes classification with sample for improvement. As a result, it improves the adaptation of the system for filtering the emerging spam. Via our test, the SpamBlocker is superior in anti-spam and it has been successfully applied in practice.The structure of this thesis is organized as follows: chapter one introduces the general status of spam, anti-spam technology and main concerns in this thesis; chapter two introduces several corresponding technology used in SpamBlocker; chapter three introduces the general design for SpamBlocker; chapter four and five respectively design and implement server and client of the MailFilter, focusing individual part更多还原