近年来,随着社交网络的快速发展,社交网络已成为僵尸网络隐匿和攻击的理想平台.僵尸网络利用社交网络作为命令与控制传播通道,通过含有控制指令或恶意程序的异常页面来传播命令和控制僵尸主机.这种攻击方式具有隐秘性高的特点,使得传统的僵尸网络检测技术的效果大打折扣.因此如何检测出含有异常文本的页面是社交僵尸网络检测面临的一个重要问题.文章将机器学习算法应用于社交网页检测中,设计并实现了一个异常页面检测系统.文章首先设计爬虫工具收集社交网络中的网页数据,然后借鉴文本分析的方法对页面进行异常特征提取,进而利用KNN和SVM分类算法对特征向量集进行判断,最后对判断结果做出评估分析.实验表明该异常页面检测系统能够有效检测异常页面,提高检测效率,为进一步发现僵尸网络提供依据.%In recent years, with the rapid development of social networks, social networks have become an ideal platform for the botnets to conceal and attack. Botnets use social networks as command and control channels, spreading commands and controlling Zombie hosts by using abnormal pages that contain the control instructions and malicious programs. This way of attack is characterized by high conifdentiality and the effects of the traditional botnet detection technologies in turn are greatly reduced. So how to detect the pages containing the abnormal texts is an important problem that the social botnet detection faces. This paper applies the machine learning algorithm to detect abnormal pages, and designs and achieves an anomaly detection system. Firstly, this paper designs crawler tool to collect Web data, then extracts the abnormal characteristics of pages in terms of the method of text analysis, and uses KNN and SVM classiifcation algorithms to determine the characteristic vectors set, ifnally gives the analysis of the judgment result. Experiment shows that the anomaly detection system can effectively detect abnormal pages and improve the detection efifciency, providing the support for ifnding botnets next step.
展开▼
