摘要
为解决日益严峻的网络钓鱼问题,提出一种基于消息传递注意力网络(Message Passing Attention Network,MPAN)的钓鱼URL检测方法。此方法相对于传统的机器学习和黑名单检测方法,无需人工提取特征且能识别新出现的钓鱼网页。首先基于敏感词分词的方法对URL分词,以提升利用URL数据信息的程度。然后通过MPAN获取URL中长距离、非连续的单词交互信息,基于自动提取的特征检测钓鱼网页。实验结果表明,基于MPAN的钓鱼网页检测方法能够达到较高的准确率、召回率、F1值。
To solve the increasingly serious problem of phishing,a phishing URL detection method based on message passing attention network(MPAN)is proposed.Compared with traditional machine learning and blacklist detection methods,this method does not need to extract features manually,and can recognize new phishing web pages.Firstly,URL is segmented based on sensitive word segmentation method to improve the degree of using URL data information.Then,the long-distance and discontinuous word interaction information in the URL is obtained through MPAN,and the phishing web page is detected based on the automatic feature extraction.Experimental results show that the method based on MPAN can achieve high accuracy,recall and F1 value.
作者
张桥
卜佑军
陈博
曹东伟
张稣荣
ZHANG Qiao;BU Youjun;CHEN Bo;CAO Dongwei;ZHANG Surong(Zhongyuan Network Security Research Institute, Zhenzhou University, Zhengzhou 450001, China;Information Engineering University, Zhengzhou 450001,China)
出处
《信息工程大学学报》
2021年第4期443-449,共7页
Journal of Information Engineering University
基金
国家重点研发计划资助项目(2017YFB0803201,2017YFB0803204,2016YFB0801200)
国家自然科学基金资助项目(61572519,61802429,61521003)
上海市科学技术委员会科研计划项目(16DZ1120503)
中国博士后基金资助项目(44595)。