基于复杂网络理论的微博用户关系网络演化模型研究

微博给人们提供便利的同时也产生了较大的负面影响.为获取微博谣言的传播规律,进而采取有效措施防控其传播,本文基于复杂网络理论研究微博用户关系网络的内部特征,提出一种微博用户关系网络演化模型,借助于平均场理论,分析该演化模型的拓扑统计特性,以及谣言在该演化模型上的传播动力学行为.理论分析和仿真实验表明,由该模型演化生成的微博用户关系网络具有无标度特性.度分布指数不仅与反向连接概率有关,而且还取决于节点的吸引度分布.研究还发现,与指数分布和均匀分布相比,当节点吸引度满足幂律分布时,稳态时的谣言传播程度较大.此外,随着反向连接概率或节点初始连边数量的增加,谣言爆发的概率以及网络中最终接受谣言的节点数量都会明显增大.

An evolution model of microblog user relationship networks based on complex network theory

Microblog provides convenience to the society, but at the same time, it also brings some adverse effects. To obtain the propagation mechanism of microblog rumor, and then take effective measures to prevent its spread, according to the complex network theory, in this paper we investigate the internal characteristics of microblog user relationship networks, and present a microblog user relationship network evolution model. By using the mean-field theory, the topological statistical property of our evolution model, and the dynamical behaviors of rumor spreading on such a model are analyzed. Theoretical analysis and simulation results show that such an evolving network exhibits a scale-free property. The degree distribution exponent not only is related to the reverse connection probability, but also depends on the node attraction degree distribution. It is also found that when the node attraction degree follows a power-law distribution, the steady-state rumor prevalence is great compared with the exponential distribution and uniform distribution. Moreover, as the reverse connection probability or the number of node initial edges increases, the probability of rumor outbreak and the number of nodes finally infected by the rumor will also increase.
Keywords： complex network mean-field theory node attraction degree rumor spreading