Studies in Science of Science ›› 2026, Vol. 44 ›› Issue (3): 661-672.

Previous Articles    

Analysis of the construction of the meta-universe "user base" from the perspective of agenda setting game:Research based on deep learning and text analysis

  

  • Received:2025-01-05 Revised:2025-03-24 Online:2026-03-15 Published:2026-03-15

议程设置视角下元宇宙“用户基础”建设研究———基于深度学习文本分析的研究

程絮森,张晓玲,薛童,雷思羽,杨波   

  1. 中国人民大学
  • 通讯作者: 杨波
  • 基金资助:
    国家自然科学基金重点项目“网约车平台数字化监管治理机制与政策研究”;国家自然科学基金面上项目“基于协作工程的AI赋能人机团队协作中的信任研究”;教育部哲学社会科学研究重大课题攻关项目“生成式人工智能助推数字经济高质量发展的机制与路径研究”

Abstract: Based on the agenda setting theory and technology acceptance model, this research utilizes deep learning and text analysis method to explore the state, causes and effective road to construct meta-universe public cognition from the two levels of information transmission and reception. By analyzing the meta-universe news and user’s comments data from 2021-2023, this research reveals the meta-cosmic agenda communication is a game involving multiple agents, and the secondary communication of information by opinion leaders dominates the formation of public emotional cognition. The content of the media agenda does not coincide with the focus of public comment and contrary to the news, the sentiment of user comments is negative, which is due to the interaction of multiple external variables such as policy culture, which affect the perception of technological usefulness and ease of use. This paper analyzes the “user element” of meta-universe from multiple angles and at a deep level, which provides reference for the new quality productivity forces integration of meta-universe innovation and development, meta-universe policy formulation, product innovation and content publicity.

摘要: 基于议程设置理论和技术接受模型,利用深度学习文本分析方法,分析2021-2023年元宇宙新闻报道和用户评论数据,从信息传播和接收两个层面探究元宇宙公众认知的现状、成因和有效建设链路。研究发现:元宇宙议程传播是多主体参与的博弈,意见领袖对信息的二次传播主导着公众情感认知的形成;内容上媒体议程与公众评论话题聚焦点并不一致;情感上二者相悖,用户评论呈负面消极情绪,原因在于政策文化等多外部变量互相作用,影响了技术有用性和易用性的感知。研究多角度、深层次地对元宇宙用户要素进行分析,为新质生产力融合元宇宙创新发展、元宇宙政策制定、产品创新及内容宣传提供参考启示。

CLC Number: