情报科学 ›› 2024, Vol. 42 ›› Issue (8): 144-153.

• 业务研究 • 上一篇    下一篇

青少年对网络信息内容治理的态度研究 ——基于B站视频弹幕和评论

  

  • 出版日期:2024-08-01 发布日期:2024-11-05

  • Online:2024-08-01 Published:2024-11-05

摘要: 【目的/意义】通过挖掘青少年对网络信息内容治理的态度,促进相关治理政策的完善和治理效果的提升, 助力健康网络环境的营造。【方法/过程】基于 CAPS理论构建青少年网络信息内容治理态度的研究思路,结合 B站 评论和弹幕,利用TF-IDF优化的LDA主题模型和共词分析法挖掘青少年治理认知,并采用ERNIE模型进行情感 分析,最后利用 IPA 建立关注度和满意度框架并提出针对性建议。【结果/结论】研究发现:①青少年的隐私意识较 强,对隐私问题治理比较关注,形成了一定认知,但忽视了网络内容记忆相关风险的治理;②青少年对于网络信息 内容治理表现出较强的爱国情怀和社会责任感;③青少年对于现阶段网络信息内容治理大多数保持中性或积极支 持的情感态度。【创新/局限】在研究网络信息内容治理态度和效果时,创新性地将青少年作为切入点;并将共词分 析法应用于不同主题关键词的关联性分析中;同时使用ERNIE提高弹幕文本分类准确性,具有一定创新性。但因 实施各治理措施的时间存在差异导致分析缺乏动态性。

Abstract: 【Purpose/significance】By exploring the attitude of young people towards online information content governance, promote the improvement of relevant governance policies and the improvement of governance effects, and help create a healthy online environ⁃ ment【. Method/process】Based on CAPS theory, the research ideas on adolescent online information content governance attitude were constructed, combined with comments and danmaku from Bilibili, the LDA theme model optimized by TF-IDF and co-word analysis was used to explore the governance cognition of adolescents, and then the ERNIE model was used for sentiment analysis, and finally IPA was used to establish the attention and satisfaction framework and put forward targeted suggestions.【Result/conclusion】It is found that: ① Adolescents have a strong sense of privacy, pay more attention to the governance of privacy issues, and form a certain cognition, but ignore the governance of risks related to online content memory; ② Adolescents show strong patriotism and sense of so⁃ cial responsibility for online information content governance; ③ Adolescents maintain a neutral or actively supportive emotional atti⁃ tude towards the current stage of online information content governance.【Innovation/limitation】It innovatively takes teenagers as the starting point in the study of the attitude and effect of Internet information content governance. In addition, it applies co-word analysis to the correlation analysis of keywords of different themes and uses ERNIE to improve the accuracy of danmaku classification, which is innovative. However, the analysis lacks dynamics.