基于计算叙事模型的虚假健康信息特征研究  

Research on the Characteristics of False Health Information Based on Computational Narrative Model

作　　者：孔杉杉 张军[1] 冯立超 Kong Shanshan;Zhang Jun;Feng Lichao(School of Management,Shandong University of Technology,Zibo 255000)

机构地区：[1]山东理工大学管理学院,淄博255000

出　　处：《情报杂志》2024年第2期152-161,共10页Journal of Intelligence

基　　金：国家自然科学基金项目“社交媒体上观点、信息和行为的耦合动力学机制研究”(编号:G71801145);山东省自然科学基金项目“网络虚假健康信息传播机制与治理策略研究”(编号:ZR2023MG032)研究成果。

摘　　要：[研究目的]面对海量健康信息,如何快速准确辨别信息真伪尤为重要。基于计算叙事模型对虚假健康信息的特征进行量化分析,为准确判别虚假健康信息提供方法支持。[研究方法]基于叙事学理论,构建社交媒体上虚假健康信息的计算叙事模型,从叙事内容、叙事逻辑和叙事链接三方面刻画虚假健康信息的特征,提出了基于语义角色标注的特征降维提取方法,从虚假健康信息文本中构建计算叙事网络,并通过网络结构分析比较不同类型虚假健康信息的特征差异。[研究结论]实证结果显示,叙事网络能刻画虚假健康信息的叙事策略和意图,并适用于不同类型虚假健康信息特征量化和分析。网络中的语义团块表示了叙事活动中的重要概念,叙事内容展示了不同健康领域中引人瞩目的问题。虚假健康信息的说服过程贯穿了行为诱导、夸张、心理认同等多种叙事逻辑,且多种逻辑之间通过结构洞和集聚型链接模式相辅相成。[Research purpose]In the face of massive amounts of health information,how to quickly and accurately discern the authenticity of the information is particularly crucial.The quantitative analysis of the characteristics of false health information is performed based on the computational narrative model to provide methodological support for accurately distinguishing false health information.[Research method]A computational narrative model of false health information on social media is built based on narratology theory,and the characteristics of false health information are characterized by three categories:narrative content,narrative logic,and narrative links.A feature dimensionality reduction extraction method based on semantic role labeling is proposed,a computational narrative network is constructed from false health information texts,and network structure analysis is used to compare the feature differences of different types of false health information.[Research conclusion]The empirical results show that the narrative network can describe the narrative strategies and intentions of false health information,and that it is appropriate for quantifying and analyzing the characteristics of various types of false health information.Semantic blobs in the network represent key concepts in narratives that highlight compelling issues in various health domains.The persuasion process of false health information involves a number of narrative logics,including behavior induction,exaggeration,and psychological identification,and the various logics complement each other through structural holes and clustered link patterns.

关 键 词：虚假健康信息 虚假信息特征 叙事模型 计算叙事 叙事逻辑 叙事链接模式 叙事学理论 

分 类 号：G209[文化科学—传播学]