智能推荐算法作用下重大突发事件群体信息从众行为研究——基于社会学习理论视角  

Research on Group Information Herd Behavior in Major Emergent Events Under the Influence of Intelligent Recommendation Algorithms:A Social Learning Theory Perspective

作　　者：乌吉斯古楞 王晰巍[1,3,4] 王楠阿雪[3] Wuji Siguleng;Wang Xiwei;Wang Nanaxue(School of Business and Management,Jilin University,Changchun 130022;School of Economics and Management,Inner Mongolia University,Hohhot 010021;Institute of National Development and Security Studies,Jilin University,Changchun 130022;Research Center for Big Data Management,Jilin University,Changchun 130022)

机构地区：[1]吉林大学商学与管理学院,长春130022 [2]内蒙古大学经济管理学院,呼和浩特010021 [3]吉林大学国家发展与安全研究院,长春130022 [4]吉林大学大数据管理研究中心,长春130022

出　　处：《图书情报工作》2024年第20期87-103,共17页Library and Information Service

基　　金：吉林省自然科学基金面上项目“重大突发事件下智能推荐算法对网络舆情演化影响及风险预警研究”(项目编号:20240101372JC)研究成果之一。

摘　　要：[目的/意义]在当前重大突发事件频发的背景下,社交媒体的关键角色突显,尽管其智能推荐算法提升了灾害信息关注度,但也可能加剧网络谣言传播和群体信息从众行为发酵,智能推荐算法作用下重大突发事件中群体信息行为影响因素,可了解群体信息跟随倾向和决策机制,制定更为有效的危机管理和应急预案。[方法/过程]结合内容编码分析与模糊集定性比较分析方法,采用先探索后验证的研究策略,设计顺序混合研究方法。基于社会学习理论识别重大突发事件情境中智能推荐算法对群体信息从众行为影响的关键因素。运用模糊集定性比较分析方法构建组态路径分析框架。[结果/结论]研究识别出重大突发事件下8个影响群体信息从众行为的关键因素,包括观察学习、感知社会影响、算法示能性、算法体验、社会价值、社会认同、自我调节和自我效能等。基于这些因素,发现出两大类3种频发信息从众行为模式,即从众机制因素驱动的群体信息从众行为、三元交互与观察模仿他人混合的群体信息从众行为,以及三元交互与低估自身信息混合的群体信息从众行为。这些发现为深入理解社交媒体在重大突发事件应对中的作用提供新的视角,并为优化特定情景下社交媒体平台中的智能推荐算法提供理论支持。[Purpose/Significance]In the current context of frequent major emergencies,the pivotal role of social media has become evident.Although its intelligent recommendation algorithms have increased the visibility of disaster information,they might also exacerbate the spread of internet rumors and the fermentation of group information herd behavior.Understanding the influencing factors of group information behavior in major emergencies under the intelligent recommendation algorithms can reveal the tendencies and decision-making mechanisms of group information following,enabling the formulation of more effective crisis management and emergency plans.[Method/Process]This paper combined content coding analysis with fuzzy-set Qualitative Comparative Analysis(fsQCA)methods,adopting an exploratory-then-confirmatory followed by a verification research strategy,and designed a sequential mixed research method.It identified the key factors influencing group information herd behavior in the context of major emergencies based on social learning theory.The fsQCA method was used to construct a configurational path analysis framework.[Result/Conclusion]The study identifies eight key factors influencing group information herd behavior during major emergencies,including observational learning,perceived social influence,recommendation algorithmic affordance,recommendation algorithm experience,perceived social value,perceived social identity,self-regulation,and self-efficacy.Based on these factors,it discovers three distinct types of information herd behavior in two categories:information herd behavior driven by herd mechanism factors,information herd behavior involved by triadic interactions and imitation of others,information herd behavior involved by triadic interactions and discount own information,and non-herd information behavior.These findings offer new perspectives for a deeper understanding of the role of social media in responding to major emergencies and provide theoretical support for optimizing intelligent recommendation alg

关 键 词：智能推荐算法 群体信息 从众行为 影响因素 社会学习理论 

分 类 号：G251[文化科学—图书馆学]