Huawei Technologies Co. Ltd., Hong Kong, China.
City University of Hong Kong, Hong Kong, China.
Sci Rep. 2017 Jun 16;7(1):3723. doi: 10.1038/s41598-017-03613-z.
This paper establishes a Markov chain model as a unified framework for describing the evolution processes in complex networks. The unique feature of the proposed model is its capability in addressing the formation mechanism that can reflect the "trichotomy" observed in degree distributions, based on which closed-form solutions can be derived. Important special cases of the proposed unified framework are those classical models, including Poisson, Exponential, Power-law distributed networks. Both simulation and experimental results demonstrate a good match of the proposed model with real datasets, showing its superiority over the classical models. Implications of the model to various applications including citation analysis, online social networks, and vehicular networks design, are also discussed in the paper.
本文建立了一个马尔可夫链模型，作为描述复杂网络演化过程的统一框架。所提出模型的独特之处在于它能够描述可以反映度分布中观察到的“三分法”的形成机制，并且可以推导出其闭式解。该统一框架的重要特例包括泊松分布、指数分布和幂律分布网络等经典模型。仿真和实验结果都表明，该模型与真实数据集具有很好的匹配度，优于经典模型。本文还讨论了该模型在引文分析、在线社交网络和车辆网络设计等各种应用中的意义。
