基于异质网络的意见领袖社区发现

摘要

针对意见领袖社区发现问题,通过将论坛中主题及其回复关系建模为异质网络,准确表示社区结构.提出意见领袖社区影响力概念及其量化方法,在此基础上设计了一种基于异质网络的意见领袖社区发现算法.通过采集天涯论坛的大量数据,验证了该社区挖掘方案能够较准确地挖掘论坛中的意见领袖社区.%In order to discover the public opinion leader community in a web forum,a heterogeneous network is constructed to represent the relationship between the topics and their replies. By quantifying the influences of forum users, a forum opinion leader community mining algorithm (FOLCM) is put forward. The algorithm is evaluated with real large-scale web data from the TianYa forum. The results demonstrate the effectiveness of the forum opinion leader community mining algorithm.