B站学习类视频的收藏数及其影响因素分析

B站视频的收藏数是衡量用户对视频内容认可程度和兴趣程度的重要指标,该文研究B站学习类视频的收藏数及其影响因素,通过收集学习类视频信息,使用LDA模型对视频标题进行分类,利用TF-IDF技术量化视频标题信息,进而采用描述性统计和逻辑回归模型来探究各变量对收藏数的影响程度。结果表明,视频学科分类对收藏数具有显著影响,其中计算机和艺体相关教程的收藏数最高,视频标签数量、视频发布时间、视频标题信息得分和视频时长对收藏数也有显著影响。

The collection number of Bilibili website video is an importantindicator to measure users'recognition and interest in video content.This paper studies the collection number of learning video on Bilibili websites and its influencing factors,LDA model was used to classify video titles,TF-IDF technology was used to quantify video title information,and then descriptive statistics and logistic regression model were used to explore the impact of each variable on the number of collections.The results show that the classification of video subjects has a significant impact on the number of collections,and the collection of computer and art-related courses is the highest,the number of video tags,video release time,video title information score and video duration also have significant impact on the number of collections.