基于Peoplerank的微博用户可信度排序算法  

User Credibility Evaluation Algorithm Based on Peoplerank

作　　者：李文政[1] 张云飞[1] 周思琪[1] 李鑫[1] 

机构地区：[1]河海大学计算机与信息学院,南京211100

出　　处：《微型电脑应用》2017年第5期4-7,共4页Microcomputer Applications

基　　金：水利部公益性行业科研专项重点项目(201501007);NSFC-广东联合基金重点项目(U1301252);国家科技支撑计划项目(2013BAB06B04;HNKJ13-H17-04);国家自然科学基金面上项目(61272543)

摘　　要：为了提升微博用户可信度排序的稳定性与合理性,结合Peoplerank的基本思想,提出了一种基于用户关注圈特征的微博用户可信度评估的算法。引入用户关注圈平衡特征作为rank值传递的影响因子,优化了用户rank值的传递过程,对每次迭代过程中的rank值的进行加权和修正。实验结果表明,该算法对不同规模的社交用户数据集具有良好的适应性,对用户可信度的排序和评估更加合理。In order to enhance the stability and rationality in ranking the credibility of micro-blog users, an attention circle feature-based user credibility evaluation algorithm is proposed with the use of the idea of Peoplerank. The algorithm introduces the characteristics of the user＇s attention circle as the key factor to weight and modify the process of iteration, then the rank value is optimized. Experimental result shows that the proposed algorithm is suitable for the social data sets in different sizes, and it is more rational to sort and evaluate the credibility of user.

关 键 词：微博 用户可信度 Peoplerank 排序 关注圈 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]