基于权重优化LSTM网络跨站脚本攻击检测的研究  

Security classification of cross-site scripting statements in cyberspace

作　　者：李鲁群[1] 徐孟达 Li Luqun;Xu Mengda(College of Information,Mechanical and Electrical Engineering,Shanghai Normal University,Shanghai201400)

机构地区：[1]上海师范大学信息与机电工程学院,上海201400

出　　处：《网络空间安全》2019年第7期12-19,共8页Cyberspace Security

摘　　要：随着互联网世界Web2.0时代的到来,网络空间的开放程度在逐渐地提高,随之而来的是PC端网络交互的安全性受到极大的威胁。在开放式Web应用安全项目基金会(OWASP)历年发布的十大最关键Web应用安全风险中,跨站脚本攻击和注入漏洞一直高居前十位。文章在跨站脚本攻击语句的文本特性描述和分词基础上,将Word2Vec和LSTM相结合,并在LSTM输入前端,对词向量引入权重优化处理,对跨站脚本攻击检测的方法进行了研究。With the advent of the era of Web2.0,the openness of cyberspace has been greatly improved.While the security of PC terminal network interaction has been greatly threatened.Among the most critical Web application security risks published over the years by the open Web application security project foundation(OWASP),cross-site scripting attacks and injection vulnerabilities have been among the top ten.Articles in cross-site scripting statements based on text characterization and participles,combining Word2Vec and LSTM,and,therefore,in LSTM input front weight vector is introduced into the optimization,the method to detect a cross-site scripting attack is studied.

关 键 词：跨站脚本攻击 Word2vec 长短时记忆神经网络 自然语言处理 WEB安全 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]