基于遗忘曲线的用户影响力时效性度量方法

发布时间：2018-12-25 13:37

【摘要】：社交网络中的用户影响力研究对于客观认识信息传播规律、分析网络成员和辅助舆情监控等具有重大意义。现有工作对影响力度量的时效性考虑不足,导致难以准确预测影响力的变化规律。针对此问题,提出一种基于遗忘曲线的用户影响力时效性度量方法,利用遗忘曲线和重复学习理论,并结合用户活跃度参数拟合影响力随时间下降及上升的变化过程。基于新浪微博和人人网的真实数据集,对该方法进行了仿真验证,结果表明:提出的方法能简便快速较准确地刻画用户影响力在一段时期内的变化情况。另外,虽然基于遗忘曲线的方法并没有实时准确预测出用户影响力的升高降低变化规律,但却能根据初始影响力和用户的活跃状态大致给出一个用户影响力的变化区间,有利于研究者对于用户影响力未来变化趋势作一个大致的预测。
[Abstract]:The research of user influence in social network is of great significance for understanding the law of information dissemination objectively, analyzing network members and assisting public opinion monitoring and so on. Due to the insufficient consideration of the timeliness of influence measurement, it is difficult to predict the law of influence change accurately. In order to solve this problem, a method of measuring the time-effectiveness of user's influence based on forgetting curve is proposed. Using the theory of forgetting curve and repeated learning, and combining with the user's activity parameter, the process of influence decreasing and rising with time is fitted. Based on the real data sets of Sina Weibo and Renren, the simulation results show that the proposed method is simple, fast and accurate to describe the change of user influence in a period of time. In addition, although the method based on forgetting curve does not accurately predict the change rule of user influence in real time, it can give a variation range of user influence according to the initial influence and the active state of user. It is helpful for researchers to make a rough prediction of the future trend of user influence.
【作者单位】：河南省农业广播电视学校;国家数字交换系统工程技术研究中心;
【基金】：国家自然科学基金青年项目(61309020);国家自然科学基金创新群体项目(61521003) 国家重点研发计划项目(2016YFB0800100,2016YFB0800101)
【分类号】：TP393.09

【参考文献】