基于支持向量机的炒作微博识别方法  

Hype Microblog Recognition Method Based on Support Vector Machine

作　　者：董雨辰 刘琰 罗军勇 张进 

机构地区：[1]数学工程与先进计算国家重点实验室,郑州450001

出　　处：《计算机工程》2015年第3期7-14,共8页Computer Engineering

基　　金：国家自然科学基金资助项目(61309007);国家"863"计划基金资助项目(2012AA012902);国家科技支撑计划基金资助项目(2012BAH47B01)

摘　　要：微博是舆论传播的中心和渠道,同时参与舆论的形成、发展与引导过程,其自媒体发布、意见领袖参与等因素在一定程度上造成了微博谣言、虚假炒作、社会动员等现象。针对炒作微博的传播特点,分析其群体的隐蔽策划现象,挖掘出普通微博和炒作微博在传播网络结构、转发增量统计等方面的差异。通过社交网站的应用程序接口对目标微博的所有评论、转发和点赞用户进行信息获取,构建该微博的传播网络,利用社团模块度、平均最短路径和网络直径这3个属性度量该网络的紧密程度,基于支持向量机对所抽取的微博进行分类,进而识别出炒作微博。实验结果表明,该方法对微博传播用户的属性信息依赖小以及传播网络结构特征敏感,并且具有较高的炒作微博识别准确率。Microblog is not only a center or channel of mass media,but also involved in the formation,development and guidance of public opinions. The propagation of speculation microblog which is released from We-media,opinion leaders or some other users,causes microblog rumors,false hype,social mobilization and other problems. This paper analyzes the phenomenon of covert planning,mines the difference of the structure in communication networks and the incremental statistics of forwardings between the ordinary and the speculation. A novel algorithm for hype microblog recognition is proposed in this paper based on Support Vector Machine（SVM）which uses the modularity peak spread and the average diameter of the shortest path in propagation network. The proposed method has advantages of less dependence on user profile information and is sensitive to the structure of propagation networks,and it has higher recognition accuracy.

关 键 词：社交网络 炒作群体 炒作微博 社团模块度 网络直径 平均最短路径 支持向量机 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]