虚拟社区中人际交互行为的统计分析研究

本文对互联网论坛中用户的交互行为进行实证分析,建立基于隐式交互行为的虚拟社区网络,通过考察其统计特性,发现该虚拟社区网络是有向、不对称、异配的无标度网络,且具有分层性质和社团结构.用户的浏览量与其入度正相关,二者均可描述用户的影响力和受关注程度.本文从定量的角度发现,论坛中具有"贡献驱动"特性的权威用户在信息传播中的作用不容小觑;网络论坛的成功需要提高大多数惰性用户的活跃度;用户要想在虚拟社区网络中占据权威地位,需要发表高质量的帖子或形成个人品牌. 关键词： 统计物理 复杂网络 人际交互 幂率分布

Empirical analysis of interpersonal interacting behavior in virtual community

Internet forum is the place where online mass incidents usually happen, and is very important in information diffusion and opinion spreading. We empirically analyze the interacting behavior of internet forum users, construct the virtual community network based on implicit interaction behavior. Through inspecting the statistical properties, we find that this virtual community network is directional, asymmetric and disassortative scale free network, and it has hierarchy and community structure.Moreover, user page view is positively correlated with his in-degree, so both can describe the influence power and popularity. From quantitative view, the results show that the effects of influential user with 'contribute-driven' property on information spreading can not be ignored; the success of internet forum needs to increase the activity of the most inert users; if user wants to be authoritative in virtual community networks, he needs to propose interesting message or has a personal brand.