一种新的微博社区发现算法

A NEW MICRO-BLOG COMMUNITY DETECTION ALGORITHM

在舆情分析、微博营销和个性化推荐等方面,微博社区发现的研究都具有重要的应用价值。为了准确而有效地发现微博社交网络中的社区,提出一种基于信任关联度的微博社区发现算法(TRKM算法)。该算法通过微博用户的评论、转发、原创微博等属性来构造节点间信任关联度,再利用微博社区的模块度对网络社区划分效果进行评价。在新浪微博明星和普通用户数据集上进行实验,并将TRKM算法与传统K-means算法作比较。实验表明,该算法能够更有效地发现微博用户关系网络中的社区结构。

The research on micro-blog community detection has important application value in public opinion analysis, microblog marketing and personalized recommendation, etc. In order to find communities in micro-blog social networks accurately and efficiently, this paper proposes a micro-blog community detection algorithm based on trust relation degree （TRKM algorithm）. This algorithm constructs the trust relation degree between the nodes through user comments, forwarding number, original micro-blog article number and other attributes, and uses the module degree of micro-blog community to evaluate the effects of network community partition. Experiments are carried out respectively on the Sina micro-blog dataset of stars and ordinary users to compare TRKM algorithm with the traditional K-means algorithm. Experimental result indicates that TRKM algorithm can more effectively find the community structure in mirco- blog user relationship networks.