摘要
针对传统Web指纹识别方法中识别对象局限于主流Web服务器软件的问题,提出一种基于余弦测度下K-means的网络空间终端设备识别模型.首先,设计识别模型和确定验证方法.其次,选取返回的HTTP数据包头部字段和状态码作为终端设备特征,对特征进行提取和向量化后转化为32维特征向量.再次,选取余弦距离函数作为K-means聚类算法中的相似性度量函数.最后,根据识别模型设计实验算法流程,对网络空间中的无标记样本和标记样本进行识别实验.实验结果表明,该模型能够识别无线路由器、网络摄像头和智能交换机等终端设备,并具有较高的识别准确率和较低的识别遗漏率.
Since the traditional web fingerprinting methods are limited to identification of mainstream web server softwares,a kind of cyberspace device identification model based on K-means with cosine distance measure is proposed. Firstly,identification model is designed and verification method is determined. Secondly,the header fields and the status code of HTTP response are selected as characteristics of terminal device and then the characteristics are transformed into 32-dimensional feature vector by feature extraction and vectorization. Thirdly,cosine distance function is selected as similarity measuring function in K-means. Finally,experiment algorithm process is designed according to the identification model and the experiments for unlabeled samples and labeledsamples are carried out. The results show that the identification model works for many kinds of terminal devices,including wireless router,web camera,and intelligent switch,and has high accuracy rate and low omission rate.
出处
《中国科学院大学学报（中英文）》
CSCD
北大核心
2016年第4期562-569,共8页
Journal of University of Chinese Academy of Sciences
基金
国家自然科学基金(61562059
61461027)资助
作者简介
通信作者，E-mail：zhaojianjun@hotmail．com
