基于PageRank和系统调用的网页安全检测模型

Webpage Security Evaluation Model Based on PageRank and System Calls

传统的基于病毒特征的网页检测方法已不能适应现在病毒产生快和变形多的特点,针对这个问题,本文提出基于Pag-eRank和系统调用的网页安全检测模型.该模型首先使用PageRank对网页进行快速分类,然后通过自动分析安全网页的系统调用参数来建立系统调用的正常访问模式,以此来检测可疑网页中病毒入侵的异常访问行为.最后实现了系统原型并进行性能测试,实验结果表明:本模型对典型恶意网页的检测率达到79.5%,效果理想,尤其在对新出现的网页病毒进行检测时明显优于传统方法.

Traditional webpage evaluation methods based on the virus characteristics can not meet the features of rapid development and quick deformation, according to this problem, a new model based on PageRank and system calls was proposed to detect virus in webpage （PRSC）. Firstly, PRSC classified the webpages by PageRank and through analyzing the arguments of system calls, and then automatically learnt and built the normal access mode to detect the abnormal access behaviors generated by malicious webpages. Finally, a prototype was implemented and the experiment results showed that it worked well and achieved 79.5 percent detection rate of typical malicious webpages and was clearly superior to the traditional methods when detecting the new web-based virus.