基于论坛信息的网络水军行为特征分析

发布时间：2018-08-08 15:59

【摘要】：近年来随着互联网的普及和深入，越来越多的网络水军活跃在各大社会媒体上，包括论坛、微博、博客、淘宝等。从事话题炒作，事件营销等危害性行为，干扰舆情的正常传播，误导舆论方向，，危害舆情安全，有必要对其行为特征进行研究。文中采用基于功能语义的信息抽取技术和相关算法采集了国内五大论坛上某个版块一段时间的发布数据；并从中提取了帐户、贴文和评论的基本属性；在对网络帐户进行人工分类的基础上，分析了水军贴文和评论的分布情况；验证了水军及其行为严重影响论坛舆论的结论。贴文和评论是水军行为的结果，本文以贴文的发表时间和评论的回复时间为研究对象，通过分析不同周期内和不同时间起点的贴文和评论的统计数量分布，来寻找水军的群体性行为特征，得到了小时周期上的典型三峰分布，提供了水军群体协同性的有效证据，从行为特征上验证网络水军是一个有组织的彼此协同的行为群体。此外本文还利用社区发现算法，在范纯龙等人提出的基于账户同帖率关系网构建的基础上进行水军社区结构分析和水军组织发现，然后利用Pajak软件构建网络水军的社区结构图，从而以可视化的形式再现网络水军的组织模式。再次从社区结构和组织发现上为水军行为活动具有协同性和组织性提供了有效证据，从组织结构上验证网络水军是一个有组织有结构的，具有高度协同性的网络组织。
[Abstract]:In recent years, with the popularization and deepening of the Internet, more and more network water forces are active in the major social media, including forums, micro-blog, blogs, Taobao and so on. It is necessary to conduct research on their behavior characteristics by disturbing the normal dissemination of public opinion, misleading public opinion and endangering public opinion. The information extraction technology based on functional semantics and related algorithms are used to collect data of a period of time on a block on the five major forums in China; and the basic attributes of accounts, posts and comments are extracted from it. On the basis of artificial classification of network accounts, the distribution of Water Army posts and comments is analyzed. The water army and its behavior seriously affect the conclusion of the forum public opinion. The paper and comments are the result of the water army's behavior. This paper takes the time of publication and the reply time of the comments as the research object. By analyzing the statistical quantity distribution of the posts and comments in different cycles and different time starting points, the paper finds the characteristics of the group behavior of the water army. The typical three peak distribution on the hour period provides the effective evidence of the association of the water army groups, and verifies that the network water army is an organized and cooperative behavior group from the behavioral characteristics.
In addition, this paper also uses the community discovery algorithm, on the basis of Fan Chunlong et al. Based on the construction of the relationship network based on the account with the relationship network construction, water army community structure analysis and water army organization discovery, and then use Pajak software to construct the network water Army Community Structure Map, thus reproducing the network water army organization pattern in the form of visualization. The discovery of community structure and organization provides effective evidence for the synergism and organization of the water army's behavior. From the organizational structure, it is proved that the network water army is a structured and highly cooperative network organization.
【学位授予单位】：沈阳航空航天大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.08

【参考文献】