基于超网络的企业微博粉丝兴趣挖掘

发布时间：2019-01-11 09:30

【摘要】：微博是由Web2.0技术发展而来的新媒体产物,它是一类信息传播与共享的社交网络平台。在日常生活中,话题的传播、热点事件的讨论和网络产品销售越来越离不开微博了。近年来,国内外的学者们开始研究微博,发表的文献数量急剧增长,同时也吸引了很多其他领域的研究者加入进来。但是,现阶段微博的研究还在发展阶段,研究方法和内容尚未成熟。由于微博的数据量日益增多和信息量急速膨胀,人们对微博信息的辨识能力也减弱。微博作为虚拟社区的代表,用户之间的交流形成一个节点众多和结构复杂的复杂系统网络。一般的网络不能够完全表示用户间、话题间的关系。因此,需要建立超网络来解决这种多种拓扑性质的问题。这样,既能完全刻画出微博中两类不同质的点,又能直接美观地呈现出来。粉丝作为一个特殊群体,常常疯狂地热爱某个事物。而微博粉丝是网络粉丝中的一种,如果他“关注”了这个微博主,就成为它的粉丝。一个微博主的粉丝越多,它发表的微博信息有可能被更多的人看到,即它的影响力越大。我们对粉丝的行为进行研究,可以提高企业的品牌形象和微博营销,还可以让企业知道用户的产品体验情况。目前,我国对粉丝行为的研究偏少,而对微博粉丝的研究就更少了。首先,本文对微博超网络的研究现状进行分析,并借鉴现有的超网络模型提出了微博话题内容子网络、粉丝子网络和面向粉丝兴趣的企业微博超网络模型三种网络结构。然后,对微博话题进行切词分词,每条微博信息提取出5个关键词,将微博话题与关键词建立边的联系。粉丝通过转发或评论参与微博话题的讨论,间接地与关键词建立关系。最后,本文使用C++语言,搭建一个平台框架来抓取新浪微博数据。数据选取的是中国移动官方微博的数据。同时,通过网络模型构建关键词关系网,并进行词频分析、中心性分析和凝聚子群分析来挖掘粉丝感兴趣的核心内容,验证了本模型的有效性。
[Abstract]:Weibo is a new media product developed from Web2.0 technology. It is a kind of social network platform for information dissemination and sharing. In daily life, the spread of topics, hot issues and network product sales more and more inseparable from Weibo. In recent years, scholars at home and abroad began to study Weibo, published a rapid increase in the number of documents, but also attracted many other fields of researchers to join. However, at present, Weibo's research is still in the stage of development, research methods and content is not yet mature. Due to Weibo's increasing amount of data and rapid expansion of information, the ability of identifying Weibo information is also weakened. Weibo is the representative of virtual community, and the communication between users forms a complex system network with many nodes and complicated structure. General network can not completely express the relationship between users and topics. Therefore, it is necessary to establish a supernetwork to solve this problem of various topological properties. In this way, can completely depict two kinds of different points in Weibo, but also directly and aesthetically present. Fans, as a special group, often madly love something. Weibo fans are one of the online fans, if he "pay attention" to the Weibo Lord, become its fans. The more followers a Weibo has, the more likely it will be to see its message, the more influential it is, the more likely it is to be seen. We study the behavior of fans, which can improve the brand image of enterprises and Weibo marketing, but also let enterprises know the user's product experience. At present, there are few researches on fan behavior in China, and even less on Weibo fans. Firstly, this paper analyzes the current research situation of Weibo supernetwork, and puts forward three kinds of network structure, such as the topic content subnet of Weibo, the fan subnetwork and the enterprise Weibo supernetwork model oriented to fans' interest, using the existing supernetwork model for reference. Then, the topic of Weibo is divided into words, and five keywords are extracted from each piece of Weibo information, and then the topic is connected with the keyword. Fans through retweets or comments to participate in the discussion of Weibo topics, indirectly to establish a relationship with key words. Finally, this paper uses C language, build a platform framework to capture Sina Weibo data. The data is selected by China Mobile official Weibo data. At the same time, the keyword relationship network is constructed through the network model, and word frequency analysis, centrality analysis and condensed subgroup analysis are carried out to mine the core content of fans' interest, which verifies the validity of this model.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：G206;F713.55

【参考文献】