微博用户粉丝演化模型的构建与实证

发布时间：2018-03-22 12:21

本文选题：微博　切入点：用户粉丝　出处：《河北大学》2012年硕士论文　论文类型：学位论文

【摘要】：2009年新浪微博的出现，拉开了微博在国内的发展序幕。微博作为一个便捷的交流工具，随着时间的推移逐渐得到了国内网民的关注与喜爱。微博的蓬勃发展也引起了众多学者的研究兴趣，他们从各个方面对微博展开了深入的研究，其研究成果为人们更透彻地了解微博提供了理论基础。但是，对于微博中用户粉丝的变化情况，却很少有学者进行探究。然而，企业进行微博营销时所考虑的主要因素就是粉丝数，因此，笔者以新浪微博为例探索了用户粉丝的变化规律，希望成为企业微博营销的参考依据。为了纵向研究用户粉丝的变化，本文将网络建模和网络演化方法引入到微博领域。首先，，重点总结了国内外学者对无标度网络模型（BA模型）的扩展性研究，并回顾了复杂网络的相关理论。接下来用树状图展示了新浪微博中的用户分类，为确定用户粉丝的度分布类型奠定基础。然后将新浪微博不同群体中的用户粉丝数输入Excel中，绘制散点图，得出用户粉丝的度分布类型——被截断的幂律型度分布。以度分布类型为依据，笔者选择BA模型为基本模型，又考虑到微博用户粉丝变化的实际影响因素，所以利用吸引力机制对基本模型进行了改进，从而构建了完整的用户粉丝演化模型。为了验证模型的正确性，笔者根据模型的具体算法设计了仿真流程，并利用Matlab软件编写了仿真程序。最后，通过对比仿真结果与实证数据，证实了粉丝演化模型的有效性。不同阶段针对不同微博群体，影响用户粉丝变化的因素不同。进入微博初期，用户粉丝的增加依赖于其本身的吸引力；随着时间的推移，新用户逐渐成为老用户，此时用户粉丝数的增加程度也会受用户原有粉丝与粉丝变化率的影响，因此本文进一步分析了个人用户和组织用户的属性以及属性变化率之间的相关性。同时，以相关性为结论扩展了用户粉丝演化模型。本文最大的创新之处在于：率先将网络建模的方法引入到微博用户粉丝的分析中，构建了符合用户粉丝变化规律的演化模型，为今后此领域的研究提供了新角度。
[Abstract]:The emergence of Sina Weibo in 2009 opened the prelude to Weibo's domestic development. Weibo served as a convenient means of communication. With the passage of time, it has gradually attracted the attention and love of the domestic netizens. Weibo's vigorous development has also aroused the research interest of many scholars. They have carried out in-depth research on Weibo from various aspects. The results provide a theoretical basis for people to have a better understanding of Weibo. However, there are few scholars to explore the changes of user fans in Weibo. The main factor of Weibo's marketing is the number of fans. Therefore, the author explores the changing law of users' fans with the example of the Sina Weibo, hoping to become the reference for the enterprise's Weibo marketing. In order to study the changes of users' fans vertically, this paper introduces network modeling and network evolution methods into Weibo field. Firstly, this paper summarizes the extensibility of scale-free network model and BA model by domestic and foreign scholars. Then, the user classification in Sina Weibo is shown with a tree chart, which lays the foundation for determining the distribution type of users' fans. Then, the number of users from different groups of Sina Weibo is input into Excel. Draw scattered plot, get the user fan degree distribution type-truncated power-law degree distribution. On the basis of degree distribution type, the author chooses BA model as the basic model, and considers the actual influence factors of Weibo user fan change, so the basic model is improved by using attraction mechanism. In order to verify the correctness of the model, the author designed the simulation flow according to the specific algorithm of the model, and compiled the simulation program by using Matlab software. Finally, The validity of the fan evolution model is verified by comparing the simulation results with the empirical data. Different stages aim at different groups of Weibo, and the factors that affect the changes of users' fans are different. In the early stage of Weibo, the increase of users' fans depends on their own attractiveness; over time, new users gradually become old users. At this time, the increase of the number of users will also be affected by the change rate of the users' original fans and fans. Therefore, this paper further analyzes the correlation between the attributes of individual users and organizational users and the rate of change of attributes. Based on the conclusion of correlation, the model of user fan evolution is extended. The biggest innovation of this paper is that the method of network modeling is first introduced into the analysis of Weibo users' fans, and an evolutionary model that accords with the law of user fans' change is constructed, which provides a new angle for the future research in this field.
【学位授予单位】：河北大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F49

【参考文献】