支持向量机在垃圾邮件过滤系统中的应用.pdfVIP

支持向量机在垃圾邮件过滤系统中的应用 摘 要 随着互联网的快速普及与发展，电子邮件逐渐成为了Internet应用最广的服 务，电子邮件成为了人们日常生活中信息交流的重要手段之一，但随之而来的垃圾 邮件也越来越猖獗。在日益泛滥的垃圾邮件给我们的互联网带来麻烦的同时，垃圾 邮件过滤技术的研究也在如火如荼的展开，但是当前的基于IP地址和基于手工规 则的过滤方法存在着误判率过高以及过滤能力有限等局限性，因此基于内容的垃圾 邮件过滤技术就成了目前研究的热点。基于统计学习理论的支持向量机，在解决有 限样本、非线性及高维模式识别问题中表现出许多特有的优势。并且支持向量机具 有其他各种过滤方法无法比拟的优点，如结构风险最小化、全局唯一解等，是邮件 过滤中很有前途的一种方法。本文旨在通过引入支持向量机方法构建垃圾邮件过滤 系统，研究支持向量机在两类数据识别方面的效果，并与现有的分类方法进行比较， 进而分析支持向量机在垃圾邮件过滤系统中的有效性和可行性。 关键词：垃圾邮件，过滤，支持向量机，分类，效果比较 II 支持向量机在垃圾邮件过滤系统中的应用 Abstract With the rapid popularization and development of the Internet, e-mail has gradually become the most widely services on the Internet and one of the important means for the exchange of information in daily life. But the spam becomes increasingly rampant. In the growing flood of junk e-mail, which causes trouble for the Internet at the same time, spam filtering technology is also progressing well. Because of the current IP address and the limitations that the filtering rules using manual methods has a very high rate of misjudgment and limited capacity, content-based spam filtering technology has become the main stream of research in this field. Based on Statistical Learning Theory, the support vector machines owns many unique advantages in resolving the limited samples, nonlinear high-dimensional pattern recognition and performance problems. Support Vector Machine possesses incomparable advantages, such as structural risk minimization, globally unique solution. It is the only promising way among all the methods of filtering e-mails. This paper aims to introduce the spam filtering systems aided by Supp