摘要
针对垃圾邮件中往往会包含指向目标地址的链接的特征,提出一种新颖的垃圾邮件过滤方法.该方法提取邮件体中的链接,将其相关特征表示成一个向量,运用机器学习技术对其进行训练分类.实验结果表明,该技术是一种快速有效的垃圾邮件过滤技术.特别地,由于在垃圾邮件中添加链接对垃圾邮件发送者来说几乎是不可避免的,该技术也被认为是非常可靠的.
As spams often contain the URLs linked to target addresses,a novel method to filter the spams is proposed.In this method,email is extracted as a vector of features associated with the in-body URLs and classify it using traditional machine learning technologies.Experimental results show that,it is an efficient spam filtering technique.Since the spam senders will inevitably add the URLs to their emails,it is also confident that the method is very reliable.
出处
《郑州大学学报（理学版）》
CAS
北大核心
2009年第2期63-67,共5页
Journal of Zhengzhou University:Natural Science Edition
基金
高等学校博士学科点专项科研基金资助项目
编号20070422107
山东省科技攻关项目
编号2007GG10001002
山东省科技攻关项目
编号2008GG10001026
关键词
垃圾邮件过滤
链接特征
支持向量机
spam filtering
URL s characteristics
Support Vector Machine
