基于弹幕评论的在线知识社区用户关注度与情感度联合分析

Joint Analysis of User Attention and Sentiment Value in Online Knowledge Communities Based on Danmaku Comments

试图分析知识社区用户弹幕评论中的关注度与情感度之间的关系,并尝试构建一种关联二者的坐标系,以此为知识视频的分类与推广提供新方法。通过获取一定数量的弹幕评论,利用Sentence-LDA模型与XLNet预训练模型对弹幕评论文本进行分析。通过抽取视频弹幕评论的主题和关键词,对其进行情感估计。根据用户对不同弹幕主题的关注度和情感度,建立用户关注度-情感度坐标系。结果表明,随着视频的播放,在线知识社区用户对视频中不同主题的关注度和情感度发生明显变化。用户对视频的不同主题有不同的情感评价,即使是对同一主题,用户也会随着视频的播放而改变情感态度。依据情感与主题的演变,可以构建新的知识视频内容标签。

This study attempts to analyze the relationship between attention and sentiment of knowledge community users in danmaku comments,and tries to construct a coordinate to associate attention and sentiment,so as to provide a new method for the classification and promotion of knowledge videos.By obtaining a certain number of danmaku comments,we use the Sentence-LDA model and XLNet pre-training model to analyze the text of danmaku comments.