Abstract: The paper investigates the formation and spread characteristics of the public opinion from the perspective of network formation and evolution. This paper studies the topological structure characteristic of WEB2.0 peer production virtual community based on WEB2.0 using complex network model. The result shows that the out-degree follows power law distribution and the in-degree follows two-stage power law distribution. It also shows that the contributors behavior with strong heterogeneity. The paper uses the improved LP link prediction method to sort and analyze the potential link on WEB2.0 virtual community. As a result, in second-order path case, the object of attention by those concerned plays a more important role than other ones. The study of the case of third-order path has reached a similar conclusion; the indirect interest and the direction have a greater impact on public opinion propagation.

Key words: WEB2.0 peer production community, complex network, spread of public opinion, power law distribution

摘要： 从网络形成演化过程的视角探讨了舆情的形成与传播特征。采用复杂网络方法对WEB2.0大众生产虚拟社区的拓扑结构特征进行分析，发现虚拟社区合作网络的出度服从幂律分布，入度服从双段幂律分布，表明虚拟社区贡献者的选择行为具有较强异质性。在静态分析基础上，进一步采用基于改进LP指标的链路预测方法对WEB2.0大众生产社区舆情传播的潜在路径链接进行排序与分析，分析结果表明，在二阶路径情况下，被关注者的关注对象往往比被关注者的被关注对象更重要，在三阶路径情况下也得出类似结论，间接关注及其方向对舆情传播具有较大影响。

关键词: WEB2.0大众生产社区, 复杂网络, 舆情传播, 幂律分布