| 芦子涵,郑中团.突发事件下基于改进 K-Shell 分解的意见领袖识别研究[J].情报工程,2023,9(1):030-042 |
| 突发事件下基于改进 K-Shell 分解的意见领袖识别研究 |
| Research on the Identification of Opinion Leaders Based on Improved K-Shell Decomposition in Emergency |
| |
| DOI：10.3772/j.issn.2095-915X.2023.01.003 |
| 中文关键词: 社会网络；突发事件；网络舆情；K-Shell 分解；意见领袖 |
| 英文关键词: Social Network; Emergency; Public Opinion; K-Shell Decomposition; Opinion Leader |
| 基金项目: |
| 作者 | 单位 | | 芦子涵 | 上海工程技术大学数理与统计学院 上海 201600 | | 郑中团 | 上海工程技术大学数理与统计学院 上海 201600 |
|
| 摘要点击次数: 882 |
| 全文下载次数: 817 |
| 中文摘要: |
| [ 目的 / 意义 ] 为实现突发事件网络舆情的精准管控，对突发事件中的意见领袖识别进行研究。[ 方法 / 过程 ] 针对 K-Shell 分解使得同一核层的节点具有相同 K-Shell 值的粗粒化分解问题，结合用户自身属性与用户交互行为对核心用户进行用户重要度因子量化，并通过重构各节点 K-Shell 值的计算方法来加以改进；在此基础上定义以转发比例为权重的相邻用户重要度贡献值，从而构建一套意见领袖的识别方法。[ 结果 / 结论 ] 以“郑州地铁 7·20 事件”为例，进行实证分析。结果表明突发事件中意见领袖主要由主流媒体与自媒体两类用户组成，且意见领袖的特征与类型随舆情生命周期的变化而变化。本文提出的意见领袖识别方法能够精确地给出意见领袖的排名，较 K-Shell 分解法识别效率更高，较社交平台传统的排序方法更具可解释性。 |
| 英文摘要: |
| [Objective/Significance] In order to realize the precise management and control of network public opinion in emergencies, the identification of opinion leaders in emergencies is studied. [Methods/Process] Aiming at the coarse-grained decomposition problem that K-Shell decomposition makes nodes in the same core layer have the same K-Shell value, the user importance factor of core users is quantified to reconstruction the calculation of K-Shell by combining the user’s own attributes and user interaction behaviors; On this basis, define the contribution value of the importance of adjacent users with the forwarding ratio as the weight, so as to construct a set of opinion leaders identification methods. [Result/Conclusion] Take “Zhengzhou Metro 7.20 Incident” as an example to conduct empirical analysis. The results show that opinion leaders in emergencies are mainly composed of mainstream media and self-media users, and the characteristics and types of opinion leaders change with the life cycle of public opinion. The opinion leader identification method proposed in this paper can accurately give the ranking of opinion leaders, which is more efficient than the K-Shell decomposition method, and more interpretable than the traditional ranking method of social platforms. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
