微博话题传播的时间网络影响力模型研究

发布时间：2018-05-03 21:34

本文选题：微博话题 + 时间网络影响力模型　；参考：《图书情报工作》2016年01期

【摘要】：[目的 /意义]基于时间网络影响力模型,研究微博话题的时变传播特性。[方法 /过程]首先构建微博话题影响力网络模型,给出影响力网络的定义、关键因素分析、模型以及网络权值的计算方法,在此基础上,基于时间网络影响力模型研究微博话题时变传播特性,利用新浪微博平台及DATAMALL的最新微博话题数据仿真分析微博话题随时间的动态传播过程以及对用户的影响力强度。[结果 /结论]微博话题随时间的动态传播过程以及对用户的影响力强度之仿真分析结果表明:约有93.3%的话题延迟在1-5小时以内,同时微博话题的影响力网络权值越高,相应的转发评论人数越多,微博话题的影响力也越大。最后将本文提出的TNIM模型与传统的影响力网络模型LDA进行对比,结果显示TNIM模型的影响力网络权值的准确性和稳定性都高于LDA模型,验证了TNIM模型的有效性。
[Abstract]:Aim / significance based on time network influence model to study the time-varying propagation characteristics of Weibo topic. [method / process] first of all, the influence network model of Weibo topic is constructed, and the definition of influence network, the analysis of key factors, the model and the calculation method of network weight are given. Based on the time-network influence model, this paper studies the time-varying propagation characteristics of Weibo topic, and analyzes the dynamic propagation process of the topic with time and the influence intensity on users by using the latest topic data of DATAMALL and the Sina Weibo platform. [results / conclusion] the simulation results of the dynamic propagation process of Weibo topics over time and the influence intensity on users show that about 93.3% of the topics are delayed within 1-5 hours, and the higher the weight of Weibo topics is, The larger the number of retweeted comments, the greater the influence of Weibo's topic. Finally, the TNIM model proposed in this paper is compared with the traditional influence network model (LDA). The results show that the TNIM model is more accurate and stable than the LDA model, which verifies the validity of the TNIM model.
【作者单位】：南昌大学管理学院;华东交通大学机电学院;北京科技大学图书馆;
【基金】：江西省科技厅2013年度科技项目“多目标动态企业联盟的随机模糊合作博弈理论研究”(项目编号:20132BAB211013)研究成果之一
【分类号】：TP393.092;G206

【共引文献】