社会网络中基于信任链的主题群组发现算法

发布时间：2019-03-22 19:40

【摘要】：针对社会网络中用户群组准确发现难题,提出了一种基于信任链的用户主题群组发现方法。该方法包括3个部分:主题空间发现、群组核心用户发现和主题群组发现。首先,给出了社会网络主题群组的相关形式化定义;然后,通过主题相关度计算发现主题空间,并给出主题空间上用户兴趣度计算方法;其次,提出原子、串联和并联信任链计算模型,并给出主题空间上的信任链计算方法;最后,分别给出主题空间发现算法、核心用户发现算法和主题群组发现算法。实验结果表明,提出的用户群组发现算法相比基于兴趣度的群组发现算法和边紧密度群组发现算法,平均准确率提升4.1%和11.3%,能够有效提升用户群组组织的准确度,在社会网络用户分类识别方面具有较好的应用价值。
[Abstract]:In order to solve the problem of accurately discovering user groups in social networks, a method of user topic group discovery based on trust chain is proposed. The method consists of three parts: topic space discovery, group core user discovery and topic group discovery. Firstly, the formal definition of topic group in social network is given, and then the topic space is found through the calculation of topic correlation degree, and the method of computing user interest degree in topic space is given. Secondly, the computing models of atomic, series and parallel trust chains are proposed, and the computing methods of trust chains in the topic space are given. Finally, the topic space discovery algorithm, the core user discovery algorithm and the topic group discovery algorithm are given respectively. The experimental results show that the average accuracy of the proposed user group discovery algorithm is 4.1% and 11.3% higher than that of the interest-based group discovery algorithm and the edge-compact density group discovery algorithm. It can effectively improve the accuracy of user group organization and has good application value in the classification and recognition of social network users.
【作者单位】：同济大学电子与信息工程学院;上海师范大学信息与机电工程学院;公安部第三研究所;
【基金】：国家自然科学基金资助项目(61103069,71171148) 上海市教委科研创新项目(13YZ052) 信息网络安全公安部重点实验室开放课题资助项目(C14602)
【分类号】：TP393.09;TP393.08

【参考文献】