基于时序超网络模型的突发事件网络舆情热点话题发现与演化

陈舒婷; 疏学明; 胡俊; 解学才; 张雷; 张伽

doi:10.16511/j.cnki.qhdxxb.2023.22.007

清华大学学报（自然科学版） >

2023 , Vol. 63 >Issue 6: 968 - 979

DOI: https://doi.org/10.16511/j.cnki.qhdxxb.2023.22.007

公共安全

基于时序超网络模型的突发事件网络舆情热点话题发现与演化

陈舒婷 ,
疏学明 ,
胡俊 ,
解学才 ,
张雷 ,
张伽

展开

1. 清华大学工程物理系, 北京 100084;
2. 城市综合应急科学北京市重点实验室, 北京 100084;
3. 北京师范大学(珠海校区) 国家安全与应急管理学院, 珠海 519087;
4. 北京师范大学应急管理部-教育部减灾与应急管理研究院, 北京 100875

陈舒婷(1998—),女,硕士研究生。

收稿日期: 2022-11-04

网络出版日期: 2023-05-12

基金资助

国家重点研发计划项目(2020YFC0833400)

收起

Discovery and evolution of hot topics of network public opinion in emergencies based on time-series supernetwork

CHEN Shuting ,
SHU Xueming ,
HU Jun ,
XIE Xuecai ,
ZHANG Lei ,
ZHANG Jia

Expand

1. Department of Engineering Physics, Tsinghua University, Beijing 100084, China;
2. Beijing Key Laboratory of City Integrated Emergency Response Science, Beijing 100084, China;
3. School of National Safety and Emergency Management, Beijing Normal University at Zhuhai, Zhuhai 519087, China;
4. Academy of Disaster Reduction and Emergency Management of Ministry of Emergency Management and Ministry of Education, Beijing Normal University, Beijing 100875, China

Received date: 2022-11-04

Online published: 2023-05-12

Fold

摘要

网络舆情安全是社会安全的重要组成部分,识别和追踪热点话题是治理突发事件网络舆情的基础。现有研究具有网络舆情事件表征不全面、对于热点话题的识别和追踪局限于语义信息等问题。该研究基于社交、内容、话题、情感4个维度构造超网络模型,并引入时间特征作为网络的连接关系,用于定量表征时序的网络舆情事件;将话题节点在超网络中的中心性及中心性变化率作为话题热度的度量指标,实现热点话题发现及演化跟踪;应用“甘肃白银马拉松”微博舆情案例对模型和指标进行验证分析。研究结果表明：该时序超网络模型能够清晰表征突发网络舆情事件,中心性及中心性变化率指标能够准确识别和跟踪热点话题,并为实时态势研判预警、舆论引导等提供指导。

关键词： 超网络; 网络舆情; 中心性; 微博话题

本文引用格式

陈舒婷 , 疏学明 , 胡俊 , 解学才 , 张雷 , 张伽 . 基于时序超网络模型的突发事件网络舆情热点话题发现与演化[J]. 清华大学学报（自然科学版）, 2023 , 63(6) : 968 -979 . DOI: 10.16511/j.cnki.qhdxxb.2023.22.007

Abstract

[Objective] Network public opinion security is an important part of social security, and identifying and tracking hot topics is the basis for the governance of online public opinion in emergencies. Existing research on the evolution of hot topics is limited by text semantics, and network public opinion text is severely limited by sparsity. Existing online public opinion event models based on complex network analysis focus on the static and isolated dimension of user characteristics, which provides an incomplete representation of online public opinion events. Meanwhile, existing online public opinion event models based on the supernetwork refer to the life-cycle theory of public opinion, divide time windows according to the development stage of public opinion, and construct them as nodes in the network. Hence, these models are limited to retrospective studies and are difficult to use.[Methods] To address the aforementioned problems, first, a supernetwork model with four levels—social, content, topic, and emotion—was constructed. This model departed from previously isolated interpretations of relevant factors of hot topics in online public opinion and effectively utilized the internal relationship between multiple public opinion elements and hot topics. The model provided a reference for mining hot topic information with complex network characteristics. Second, considering the universality and timeliness of the model, time information was constructed differently from traditional research, which was based on the development stage of public opinion. In this study, time windows were divided equidistantly with a certain granularity, and the order of time windows, rather than the nodes in the network, represented the connectivity characteristics of the network. To discover, migrate, and predict hot topics, a topic popularity index based on the centrality of the supernetwork and a topic popularity change rate index based on the hypernetwork centrality change rate were proposed in this paper. These indices were verified and analyzed for the “Gansu Baiyin Marathon” Microblog public opinion event.[Results] The findings of this study are as follows: (1) The time-series supernetwork model clearly represents network public opinion events and has significant advantages over the traditional methods in model visualization. (2) The topic popularity index accurately identifies the hot topics in each time window and evaluates the changes in topic popularity throughout the development of the event. For example, “accident notification” was the most popular topic in the early stage of the public opinion event, and its heat decreased with fluctuation thereafter. The “event guarantee” topic remained popular throughout the development of the event, and its popularity fluctuated on a daily basis. (3) Based on the topic popularity curve and network structure, topics with similar communities are found to migrate, such as “liability compensation”, “competition guarantee”, and “popular science knowledge”. (4) The topic popularity change rate index effectively predicts hot topics in the next time window.[Conclusions] This paper provides a general time-series model for network public opinion events with high sparsity and complex network characteristics. Topic heat and heat change rate indices can lead to the accurate identification of hot topics as well as the accurate tracking of the evolution, migration, and prediction of hot topics. Further, this study provides intuitive and useful guidance for the governance of online public opinion in real-world situations.

Key words： supernetwork; network public opinion; centrality; Microblog topic

参考文献

[1] 刘毅.略论网络舆情的概念、特点、表达与传播[J].理论界, 2007(1):11-12. LIU Y. The concept, characteristics, expression and dissemination of online public opinion[J]. Theory Horizon, 2007(1):11-12.(in Chinese)
[2] 郭乐天.互联网虚假信息的控制与网络舆情的引导[J].新闻记者, 2005(2):23-26. GUO L T. The control of Internet false information and the guidance of Internet public opinion[J]. Shanghai Journalism Review, 2005(2):23-26.(in Chinese)
[3] GRIFFITHS T L, STEYVERS M. Finding scientific topics[J]. Proceedings of the National Academy of Sciences of the United States of America, 2004, 101(S1):5228-5235.
[4] 胡艳丽,白亮,张维明.网络舆情中一种基于OLDA的在线话题演化方法[J].国防科技大学学报, 2012, 34(1):150-154. HU Y L, BAI L, ZHANG W M. OLDA-based method for online topic evolution in network public opinion analysis[J]. Journal of National University of Defense Technology, 2012, 34(1):150-154.(in Chinese)
[5] WANG X R, MCCALLUM A. Topics over time:A non-Markov continuous-time model of topical trends[C]//Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Philadelphia, USA:ACM, 2006:424-433.
[6] 单斌,李芳.基于LDA话题演化研究方法综述[J].中文信息学报, 2010, 24(6):43-49, 68. SHAN B, LI F. A survey of topic evolution based on LDA[J]. Journal of Chinese Information Processing, 2010, 24(6):43-49, 68.(in Chinese)
[7] 周沧琦,赵千川,卢文博.基于兴趣变化的微博用户转发行为建模[J].清华大学学报(自然科学版), 2015, 55(11):1163-1170. ZHOU C Q, ZHAO Q C, LU W B. Modeling of the forwarding behavior in microblogging with adaptive interest[J]. Journal of Tsinghua University (Science and Technology), 2015, 55(11):1163-1170.(in Chinese)
[8] 陈安滢,朱昊然,苏国锋.微博用户的应急预警信息传播行为研究[J].清华大学学报(自然科学版), 2021, 61(6):527-535. CHEN A Y, ZHU H R, SU G F. Emergency warning information repost behavior of Weibo users[J]. Journal of Tsinghua University (Science and Technology), 2021, 61(6):527-535.(in Chinese)
[9] 康伟.基于SNA的突发事件网络舆情关键节点识别:以"7·23动车事故"为例[J].公共管理学报, 2012, 9(3):101-111. KANG W. Analysis of the key nodes in public opinion spread during emergencies based on social network theory:A case study of the 7·23 Wenzhou high-speed train collision[J]. Journal of Public Management, 2012, 9(3):101-111.(in Chinese)
[10] 赵宇,黄开枝,郭云飞,等.在线社会网络中面向节点影响力的信息传播阻断模型[J].清华大学学报(自然科学版), 2017, 57(12):1245-1253. ZHAO Y, HUANG K Z, GUO Y F, et al. Information diffusion blocking model of node influence-oriented in online social network[J]. Journal of Tsinghua University (Science and Technology), 2017, 57(12):1245-1253.(in Chinese)
[11] 赵剑华,万克文.基于信息传播模型-SIR传染病模型的社交网络舆情传播动力学模型研究[J].情报科学, 2017, 35(12):34-38. ZHAO J H, WAN K W. Research on the communication dynamics model of social network public opinion based on the SIR model[J]. Information Science, 2017, 35(12):34-38.(in Chinese)
[12] 范纯龙,宋会敏,丁国辉.一种改进的SEIR网络谣言传播模型研究[J].情报杂志, 2017, 36(3):86-91. FAN C L, SONG H M, DING G H. Research on an improved SEIR network rumor propagation model[J]. Journal of Intelligence, 2017, 36(3):86-91.(in Chinese)
[13] ZHAO L J, QIU X Y, WANG X L, et al. Rumor spreading model considering forgetting and remembering mechanisms in inhomogeneous networks[J]. Physica A:Statistical Mechanics and Its Applications, 2013, 392(4):987-994.
[14] 陈昱,张慧琳.社会计算在信息安全中的应用[J].清华大学学报(自然科学版), 2011, 51(10):1323-1328. CHEN Y, ZHANG H L. Overview of social computing in information security[J]. Journal of Tsinghua University (Science and Technology), 2011, 51(10):1323-1328.(in Chinese)
[15] 田儒雅,孙巍,吴蕾,等.基于超网络的图书情报领域知识合作特征分析[J].情报理论与实践, 2016, 39(10):25-30. TIAN R Y, SUN W, WU L, et al. Feature analysis of knowledge cooperation in the field of library and information science based on supernetwork[J]. Information Studies:Theory&Application, 2016, 39(10):25-30.(in Chinese)
[16] NAGURNEY A, DONG J. Supernetworks:Decision-making for the information age[M]. Cheltenham:Edward Elgar Publishing, 2002.
[17] 尚艳超,王恒山,王艳灵.基于微博上信息传播的超网络模型[J].技术与创新管理, 2012, 33(2):175-178. SHANG Y C, WANG H S, WANG Y L. The supernetwork model of information transmission based on the microblog[J]. Technology and Innovation Management, 2012, 33(2):175-178.(in Chinese)
[18] 潘芳,鲍雨亭.基于超网络的微博反腐舆情研究[J].情报杂志, 2014, 33(8):173-177, 172. PAN F, BAO Y T. Study of microblog anti-corruption public opinion based on the supernetwork[J]. Journal of Intelligence, 2014, 33(8):173-177, 172.(in Chinese)
[19] 马宁,刘怡君.基于超网络的舆论领袖识别应用研究[J].中国科学院院刊, 2012, 27(5):586-594. MA N, LIU Y J. Recognition of online opinion leaders based on supernetwork analysis[J]. Bulletin of Chinese Academy of Sciences, 2012, 27(5):586-594.(in Chinese)
[20] 武澎,王恒山,李煜,等.突发事件信息传播超网络中重要调控节点的判定研究[J].图书情报工作, 2013, 57(1):112-116, 148. WU P, WANG H S, LI Y, et al. Judgment of important control nodes in the information dissemination super-network of emergency[J]. Library and Information Service, 2013, 57(1):112-116, 148.(in Chinese)
[21] 孙英苹,刘凤鸣,高艺畅,等.基于回音室超网络的网络舆情差异化引导研究[J].情报理论与实践, 2022, 45(12):138-145. SUN Y P, LIU F M, GAO Y C, et al. Research on differentiated guidance of network public opinion based on echo chamber super-network[J]. Information Studies:Theory&Application, 2022, 45(12):138-145.(in Chinese)
[22] 刘怡君,李倩倩,田儒雅,等.基于超网络的社会舆论形成及应用研究[J].中国科学院院刊, 2012, 27(5):560-568. LIU Y J, LI Q Q, TIAN R Y, et al. Formation and application of public opinion based on supernetwork analysis[J]. Bulletin of Chinese Academy of Sciences, 2012, 27(5):560-568.(in Chinese)
[23] 崔鹏,张巍,何毅,等.突发公共事件网络舆情演化及政府应对能力研究[J].现代情报, 2018, 38(2):75-83, 95. CUI P, ZHANG W, HE Y, et al. Dynamic evolution research on the government's response capability to the public opinions in the context of public emergencies[J]. Journal of Modern Information, 2018, 38(2):75-83, 95.(in Chinese)
[24] 张连峰,周红磊,王丹,等.基于超网络理论的微博舆情关键节点挖掘[J].情报学报, 2019, 38(12):1286-1296. ZHANG L F, ZHOU H L, WANG D, et al. Key node mining of Weibo public opinion based on super network theory[J]. Journal of the China Society for Scientific and Technical Information, 2019, 38(12):1286-1296.(in Chinese)
[25] 梁晓贺,田儒雅,吴蕾,等.基于超网络的微博相似度及其在微博舆情主题发现中的应用[J].图书情报工作, 2020, 64(11):77-86. LIANG X H, TIAN R Y, WU L, et al. Microblog similarity based on super network and its application in microblog public opinion topic detection[J]. Library and Information Service, 2020, 64(11):77-86.(in Chinese)
[26] 王丹,张海涛,刘雅姝,等.微博舆情关键节点情感倾向分析及思想引领研究[J].图书情报工作, 2019, 63(4):15-22. WANG D, ZHANG H T, LIU Y S, et al. Sentiment analysis and ideological guidance of key nodes in micro-blog public opinion[J]. Library and Information Service, 2019, 63(4):15-22.(in Chinese)
[27] 刘军.社会网络分析导论[M].北京:社会科学文献出版社, 2004. LIU J. An introduction to social network analysis[M]. Beijing:Social Sciences Academic Press, 2004.(in Chinese)
[28] 平亮,宗利永.基于社会网络中心性分析的微博信息传播研究:以SINA微博为例[J].图书情报知识, 2010(6):92-97. PING L, ZONG L Y. Research on microblog information dissemination based on SNA centrality analysis:A case study with SINA microblog[J]. Document, Information&Knowledge, 2010(6):92-97.(in Chinese)
[29] FREEMAN L C. Centrality in social networks conceptual clarification[J]. Social Networks, 1978-1979, 1(3):215-239.
[30] 任晓龙,吕琳媛.网络重要节点排序方法综述[J].科学通报, 2014, 59(13):1175-1197. REN X L, Lü L Y. Review of ranking nodes in complex networks[J]. Chinese Science Bulletin, 2014, 59(13):1175-1197.(in Chinese)
[31] BLEI D M, NG A Y, JORDAN M I, et al. Latent Dirichlet allocation[J]. The Journal of Machine Learning Research, 2003, 3:993-1022.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献

访问统计