联系我们

论文查询结果	返回搜索

论文编号：	12648
作者编号：	2120192912
上传时间：	2021/6/18 13:11:13
中文题目：	基于多重社交网络的突发公共卫生事件意见领袖识别及演化研究
英文题目：	Research on the Identification and Evolution of Opinion Leaders in Public Health Emergencies Based on Multiple Social Networks
指导老师：	徐曼
中文关键字：	突发公共卫生事件；多重社会网络；意见领袖
英文关键字：	Public health emergency; Multiple social networks; Opinion leader
中文摘要：	互联网的发展改变了人们获取、传播信息的方式。社交媒体已成为重要的信息传播载体。社交媒体中一条信息是否被重要的节点，即意见领袖转发，很大程度上会影响信息的传播范围。突发公共卫生事件舆情具有复杂性、变化性、聚焦性的特点，如何在复杂的舆情环境中快速准确地识别出多重社交网络中的意见领袖，发挥其影响力缓解民众情绪，引导网民正确应对疫情，是本文的研究重点。 微博平台的社交属性积累了丰富的信息交互行为数据，多重社交网络理论有助于意见领袖的识别。本研究基于多重社交网络理论，根据微博信息交互行为特性，融合转发与评论两层信息交互行为网络，形成微博多重社会网络模型。根据网络特性设计出意见领袖挖掘算法Multi-PageRank，该算法改进了传统PageRank算法投票时按照出度评分PR值的问题，引入用户吸引度和用户影响力两个指标调节投票权值。经过与多个算法对比验证，发现该算法能够在原有识别基础上挖掘出新的意见领袖，且MPR值粒度更精细，有助于比较排名相近的意见领袖。 根据新冠肺炎确诊人数以及传染病流行程度，划分了7个时间窗口。使用Multi-PageRank算法挖掘各周期意见领袖，探究各窗口意见领袖的特征以及演化规律。根据意见领袖的演化特征总结出三类演化规律：稳定型、跳跃型以及偶发型。通过进一步分析，发现时间-话题-意见领袖间具有强关联性，随着时间推移，网民关注话题发生转移，话题下的意见领袖也就产生了演化。三类意见领袖的影响力演化规律也表明，突发公共卫生事件舆情系统经历了从无序到有序的过程，官方媒体类、新闻媒体类意见领袖在7个窗口中的影响力十分接近，呈现博弈演化的特征。最终根据前文研究结论，为突发公共卫生事件中舆情治理策略、各类意见领袖的发言策略以及意见领袖影响力的提升提出建议。
英文摘要：	The development of the Internet has changed the way people get and disseminate information. Social media has become an important vehicle for information dissemination. Whether a piece of information in social media is reposted by an important node, namely opinion leaders, will greatly affect the spread range of information. Public opinion of public health emergencies is characterized by complexity, variability and focused. The research of this paper is how to quickly and accurately identify opinion leaders in multiple social networks, exert their influence to alleviate public emotions and guide them to correctly deal with the epidemic situation. The social attribute of Weibo platform accumulates rich information interaction behavior data. The theory of multiple social networks can not only represent the interaction relationship between nodes in the real society more completely and accurately, but also connect the isolated data fragments in the social networks at all levels, which is helpful for the identification of opinion leaders. This study is based on the theory of multiple social networks. According to the characteristics of microblog information interaction behavior, the two-layer information interaction network of reposting and comment is fused to form the microblog multiple social network model. The network is a single-layer directed weighted network. Multi-PageRank, an opinion leader mining algorithm, is designed according to the characteristics of the network. This algorithm improves the problem of the traditional PageRank algorithm to score the PR value according to the output when voting, and introduces two indexes of user attraction and user influence to adjust the voting power value. Through comparison and verification with several algorithms, it is found that this algorithm can mine new opinion leaders on the basis of original recognition, and the granularity of MPR value is more fine. According to the number of confirmed cases of new coronary pneumonia and the prevalence of infectious diseases, seven time windows are divided. Multi-PageRank algorithm is used to mine opinion leaders in each cycle, and the characteristics and evolution rules of opinion leaders in each cycle are explored. According to the evolution characteristics of opinion leaders, three types of evolution laws are summarized: stable, jumping and occasional. Through further analysis, it is found that there is a strong correlation between time, topic and opinion leader. As time goes by, netizens' attention to topics shifts, and the opinion leaders under the topic also evolve. The evolution of the influence of the three types of opinion leaders also shows that the public opinion system of public health emergencies has undergone a process from disorder to order. The influence of the official media and news media opinion leaders in the seven windows is very close. Presents the characteristics of game evolution. Finally, based on the previous research conclusions, suggestions are made for public opinion governance strategies in public health emergencies, the speech strategies of various opinion leaders, and the improvement of the influence of opinion leaders.
查看全文：	预览  下载(下载需要进行登录)