一种基于声誉的不良网页协作检测机制  

A Collaborative, Reputation-Enhanced Mechanism for Malicious Web Page Detection

作　　者：李清华[1] 徐明伟[1] 

机构地区：[1]清华大学计算机科学与技术系,北京100084

出　　处：《微计算机应用》2007年第12期1250-1255,共6页Microcomputer Applications

基　　金：国家自然科学基金项目(No.60373010;No.90604024);教育部科学技术研究重点项目(No.106012);新世纪优秀人才支持计划资助

摘　　要：大量不良网页的出现威胁了互联网安全。为了检测不良网页,提出了一种基于声誉的协作检测机制,WebShield。用户对网页评分,服务器汇聚用户评分得到网页的综合评分,并依此检测不良网页。设计了一种基于贡献的声誉模型,通过用户声誉加强评分汇聚的准确性和鲁棒性。模拟表明,WebShield对不良网页的检测效率较高、对恶意用户的抵抗性较好。The emerging of a large amount of malicious web pages has made web browsing a big threat to Internet security. This paper puts forward a mechanism named WebShield to effectively detect malicious web pages through user collaborations. In WebShield, users rate pages, and their ratings are aggregated. A contribution - based reputation model is proposed to enhance the robustness of rating aggregation. Simulations show that WebShield is high in detecting accuracy and tolerant to malicious users.

关 键 词：WEB安全 不良网页 协作检测 声誉 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]