基于文本挖掘的财经视频评论分析与热度研究——以哔哩哔哩财经类视频在线评论为例
Analysis and Research on the Popularity of Financial Video Comments Based on Text Mining—Taking Online Comments on Financial Videos on Bilibili as an Example

随着数字经济与新媒体的深度融合，短视频平台已成为公众获取财经资讯和参与投资讨论的重要渠道。本研究以哔哩哔哩(Bilibili)财经类视频的在线评论为研究对象，旨在通过文本挖掘技术分析财经视频评论的内容特征和用户情感倾向，并构建视频热度评估模型。研究方法包括LDA主题模型、SnowNLP情感分析和TF-IDF特征提取技术。通过对高质量评论数据的系统分析，本研究揭示了财经内容传播的影响机制，并为创作者优化内容生产、平台完善推荐算法、投资者把握市场情绪提供了实证依据。研究结果表明，评论中积极情绪占主导，用户对财经领域整体持乐观态度，但对具体问题和潜在风险保持警惕。不同主题下的情感倾向差异显著，高频词分析揭示了用户讨论财经话题时的双重心理。
With the deep integration of the digital economy and new media, short video platforms have become an important channel for the public to obtain financial information and participate in investment discussions. This study focuses on the online comments of Bilibili financial videos, aiming to analyze the content characteristics and user emotional tendencies of financial video comments through text mining techniques, and construct a video popularity evaluation model. The research methods include the LDA topic model, SnowNLP sentiment analysis, and TF-IDF feature extraction technology. Through systematic analysis of high-quality review data, this study reveals the impact mechanism of financial content dissemination and provides empirical evidence for creators to optimize content production, platforms to improve recommendation algorithms, and investors to grasp market sentiment. The research results indicate that positive emotions dominate in comments, and users hold an optimistic attitude towards the overall financial field, but remain vigilant about specific issues and potential risks. There are significant differences in emotional tendencies under different themes, and high-frequency word analysis reveals the dual psychology of users when discussing financial topics.