基于文档对象模型结构聚类的钓鱼网页检测方法  

A Detection Method for Phishing Webpage Based on Document Object Model Structure Clustering

作　　者：冯健[1] 张莹 FENG Jian;ZHANG Ying(College of Computer Science and Technology,Xi’an University of Science and Technology,Xi’an 710054,China)

机构地区：[1]西安科技大学计算机科学与技术学院,西安710054

出　　处：《科学技术与工程》2018年第23期81-89,共9页Science Technology and Engineering

基　　金：陕西省自然科学基金(2017JQ6053)资助

摘　　要：为了解决现有的钓鱼网页分析方法,往往基于页面的文本特征,而忽略了页面的结构特征的问题,提出基于文档对象模型(document object model,DOM)结构聚类的钓鱼检测方法,其关键在于如何快速有效地计算网页的相似度。首先对获取的页面进行DOM结构解析,构建DOM树层次标签向量以刻画网页的结构特征;然后重新定义DOM树距离的概念,通过不同DOM树之间的距离来度量网页间的相似度;最后采用划分聚类思想实现网页的聚类。一系列的仿真实验表明,方法具有较高的召回率与精确率,运行时间也较短。In order to solve the problem that existing phishing webpage analysis methods often base on text features but ignore structural features of webpage,a phishing detection method of clustering based on DOM(document object model)structure is proposed.The key point is how to calculate similarities among webpages quickly and effectively.Firstly,DOM tree of the obtained webpage is parsed to construct tag vector in DOM tree hierarchy to describe the structural feature of the webpage.Then the concept of distance between DOM trees is redefined and the similarity between webpages is measured by the distance among different DOM trees.Finally,partition based method is used to implement clustering among webpages.A series of simulation experiments show that the method proposed has higher recall,precision,and shorter runtime.

关 键 词：钓鱼网页 DOM树 层次 聚类 

分 类 号：TP393.06[自动化与计算机技术—计算机应用技术]