摘要
在深入分析了传统垃圾邮件过滤技术不足的基础上,提出并实现了一种新型的基于URL过滤的垃圾邮件过滤技术(URLbased spamfiltering,UBSF)。该方法通过对比从到来邮件中提取的URL与URL库中存储的URL信息的相似性来判定垃圾邮件。通过语料库以及构建实际系统原型的测试,表明该方法具有准确性高、误报率低以及实时处理速度快的优点。
A novel anti-spam technique was proposed and implemented based on URL filtering on the basis of analysed over the traditional anti-spam techniques. The technique determined spam by comparing the similarity of the abstracted URLs from incoming mail and the URLs from the preset URL library. Experimental evaluation from both mail corpus and the constructed prototype based on the novel technique illustrates that the technique has some advantages, such as high accuracy, low false positives and good real-time processing ability.
出处
《计算机应用研究》
CSCD
北大核心
2008年第5期1537-1539,共3页
Application Research of Computers
基金
国家“973”计划资助项目(2004CB318109)