微博双向“关注”网络节点中心性及传播影响力的分析

发布时间：2018-11-22 14:02

【摘要】：根据新浪微博的实际数据,建立了两个基于双向"关注"的用户关系网络,通过分析网络拓扑统计特征,发现二者均具有小世界、无标度特征.通过对节点度、紧密度、介数和k-core四个网络中心性指标进行实证分析,发现节点度服从分段幂率分布;介数相比其他中心性指标差异性最为显著;两个网络均具有明显的层次性,但不是所有度值大的节点核数也大;全局范围内各中心性指标之间存在着较强的相关性,但在度值较大的节点群这种相关性明显减弱.此外,借助基于传染病动力学的SIR信息传播模型来分析四种指标在刻画节点传播能力方面的差异性,仿真结果表明,选择具有不同中心性指标的初始传播节点,对信息传播速度和范围均具有不同影响;紧密度和k-core较其他指标可以更加准确地描述节点在信息传播中所处的网络核心位置,这有助于识别信息传播拓扑网络中的关键节点.
[Abstract]:According to the actual data of Sina Weibo, two user relationship networks based on bi-directional "concern" are established. By analyzing the statistical characteristics of network topology, it is found that both of them have small world and no scale characteristics. Through the empirical analysis of the four network centrality indexes of node degree, compactness, medium and k-core, it is found that the node degree is distributed according to the piecewise power ratio, and the difference of the index is the most significant compared with other central indexes. Both networks have obvious hierarchy, but not all the nodes with large degrees of kernel number, and there is a strong correlation between the central indicators in the global scope, but in the larger degree of the node group this correlation is obviously weakened. In addition, the SIR information transmission model based on infectious disease dynamics is used to analyze the differences of the four indexes in describing the transmission ability of nodes. The simulation results show that the initial transmission nodes with different central indicators are selected. It has different influence on the speed and scope of information dissemination; Compactness and k-core can more accurately describe the network core position of nodes in information transmission than other indicators, which is helpful to identify the key nodes in the information dissemination topology network.
【作者单位】：北京交通大学通信与信息系统北京市重点实验室;中国科学院计算机网络信息中心;
【基金】：国家自然科学基金(批准号:61172072,61271308) 北京市自然科学基金(批准号:11DA1454) 中央高校基本科研业务费专项资金(批准号:2011YJS215)资助~~
【分类号】：TP393.092

【参考文献】