一种话题相关的微博意见领袖挖掘算法

An Algorithm of Topic-related Microblogging Opinion Leaders Mining

微博中的意见领袖对信息的快速传播起着关键作用,能在短时间内对数量众多的用户产生直接或间接的影响.在微博中,意见领袖除了具有自身属性和网络结构特征外,还与参与的话题高度相关.针对已有挖掘研究只考虑了意见领袖的局部特征以及忽略了话题相关性的问题,提出一种话题相关的意见领袖挖掘算法.该算法首先根据微博用户的自身属性及用户间话题相关的交互信息构建带权的话题相关的微博图模型,并采用随机游走的思想来寻找图模型的中心节点,以此挖掘微博中的意见领袖.在新浪微博三个话题数据集上的实验结果表明,该算法挖掘的意见领袖在扩展核心率指标上优于类似算法.

Abstraet：Microblogging opinion leaders play a key role in the rapid spread of information. They have direct or indirect impact on large numbers of users of microblogging within a short time. In microblogging, opinion leaders don＇t only stand out in their own attributes and network structure features,they are also highly relevant to the topics in which they participate. Existing researches about mi- croblogging opinion leaders mining only consider partial features and ignore the topic relevance of opinion leaders. To address this is- sue, an algorithm of topic-related opinion leaders mining was presented. First, a topic-related graph model of microblogging was built according to users＇ attributes and mutual information among them. Then the idea of random walk was adopted in the algorithm to mine microblogging opinion leaders by finding central nodes of the graph model. Experiments based on the data of three topic-related datasets of Sina MicroBlog were performed. The results show that the opinion leaders found by this algorithm perform better than similar algorithms under the criteria of ExtendedCoreRadio. Key words： opinion leader; microblogging; topic; random walk