基于HITS算法的微博用户可信度评估

发布时间：2018-03-16 09:01

本文选题：HITS算法　切入点：微博用户　出处：《山东大学学报(工学版)》2016年05期 　论文类型：期刊论文

【摘要】：以新浪微博为研究平台,在HITS(hyperlink-induced topic search)算法的基础上,提出融合用户交互行为和博文内容的微博用户可信度评估算法。分别构建基于交互行为和基于博文内容的微博用户有向链接图,图中节点表示用户,有向边体现用户基于交互行为或基于内容的指向关系;依据HITS算法计算两种拓扑结构下微博用户的权威度和中心度;以融合的权威度作为度量评估用户可信度。试验采用从新浪微博采集的数据作为测试集合,通过反复训练法获得可信度阈值,绘制不同可信度算法的用户可信度曲线,验证了算法的可行性和有效性。
[Abstract]:Taking Weibo of Sina as the research platform, on the basis of HITS(hyperlink-induced topic search algorithm, This paper proposes an algorithm for evaluating Weibo's user credibility, which combines user interaction behavior and blog content, respectively, and constructs a directed link graph based on interactive behavior and blog content, in which nodes represent the user. The directed edge reflects the user's interactive behavior or content-based pointing relationship, and calculates the authority and centrality of Weibo users under two topologies according to the HITS algorithm. In the experiment, the data collected from Sina Weibo was used as the test set, and the confidence threshold was obtained by repeated training method, and the user credibility curve of different credibility algorithms was drawn. The feasibility and effectiveness of the algorithm are verified.
【作者单位】：天津大学管理与经济学部;河北大学管理学院;河北大学计算机科学与技术学院;
【基金】：河北省社会科学基金资助项目(HB15TQ013)
【分类号】：TP393.092

【相似文献】