| 基于主题细分的社交网络用户间交互特征分析 |
| |
| 引用本文： | 杨欣谊,朱恒民,魏静,陈文.基于主题细分的社交网络用户间交互特征分析[J].情报杂志,2021(1):178-183. |
| |
| 作者姓名： | 杨欣谊 朱恒民 魏静 陈文 |
| |
| 作者单位： | 南京邮电大学管理学院;江苏高校哲学社会科学重点研究基地——信息产业融合创新与应急管理研究中心 |
| |
| 基金项目： | 国家自然科学基金项目“基于主路径网络的舆情传播态势预测与干预研究——以社会化媒体中舆情为对象”(编号:71874088)研究成果之一;国家自然科学基金项目“基于阈值的群体事件互联网舆情传递链路预测及监控机制研究”(编号:71704085)研究成果之一;江苏高校哲学社会科学优秀创新团队培育点基地“应急决策与舆情管理”(编号:2017ZSTD022)的研究成果之一;江苏省研究生科研与实践创新计划项目“江苏省23个全国百强县市创新效率评价及改进路径研究”(编号:KYCX19_0990)的研究成果之一。 |
| |
| 摘 要： | 目的/意义]针对一微博子网,从主题细分的角度对用户间历史交互记录进行研究,发现用户间交互的主题偏好特征,以期从微观层面了解用户信息传播行为的规律。方法/过程]通过用户实例分析得出对用户间交互进行主题细分的必要性;利用主题模型(LDA)对用户间历史交互记录进行主题细分,采用多维向量表示用户间在不同主题下的交互强度;通过统计分析和网络分析方法探索用户间交互的主题特征。结果/结论]各主题下用户间交互强度的分布具有长尾特征;用户间的交互内容在时序上具有主题相关性;基于多维的用户间交互强度,可抽取出特定主题下的用户交互子网。用户间交互在时序上具有主题相关性这一特征,以及特定主题的用户交互子网,可用于对特定主题的信息传播进行监控和预测。
|
| 关 键 词： | 在线社交网络 主题细分 用户间交互 时序相关性 |
Analysis of Topic-based Users'Interaction in Online Social Networks |
| |
| Yang Xinyi,Zhu Hengmin,Wei Jing,Chen Wen.Analysis of Topic-based Users'Interaction in Online Social Networks[J].Journal of Information,2021(1):178-183. |
| |
| Authors: | Yang Xinyi Zhu Hengmin Wei Jing Chen Wen |
| |
| Institution: | (School of Management,Nanjing University of Posts and Telecommunications,Nanjing 210003;Jiangsu University Philosophy and Social Science Key Research Base——Information Industry Integration Innovation and Emergency Management Research Center,Nanjing 210003) |
| |
| Abstract: | Purpose/Significance]Aiming at a sub-network from Sina Weibo,from the perspective of topic subdivison,we study the topic preference of interaction between users from the historical interactive records,which helps to understand the law of user's spreading behavior from the micro level.Method/Process]Through the analysis of several user examples we find that it is necessary to introduce the idea of topic-oriented to describe the interaction between users;with the topic model(LDA),we allocate topic vectors to the historical interaction records between users from Sina Weibo,then use multidimensional vectors to represent the strength of interaction between users;through statistical analysis and network analysis we explore the characteristics of the relationship between users under different topic.Result/Conclusion]The distribution of users'interactive relationships under each topic has long tail characteristics;the interactive content between users is similar in different time periods and can be used for predicting users'spreading behavior;based on the topic-oriented interactive relationship,the subnets of users under specific topics can be extracted.The temporal correlation of interactive topics between users,as well as the subnet under specific topics can be used to monitor and predict the information diffusion under specific topics. |
| |
| Keywords: | online social networks topic-oriented interactions between users temporal correlation |
| 本文献已被 维普 等数据库收录！ |
