Scholarlink

With the continuous development of Internet technology, evaluation video has become an important marketing tool for merchants and brands in all walks of life, and plays an important role in the purchase intention of consumers. The purpose of this study is to use deep learning technology to conduct a binary task of willing or unwilling to buy for evaluation video comments on Bilibili platform. Firstly, by using a suitable web crawler tool, all the comments are obtained from the target video, and then the necessary pre-processing work is carried out on these comments; Then, a deep learning model with strong expression ability is constructed by using neural network model, and the feature extraction of keywords in comments is carried out by CNN to capture important information in comments. Finally, through the full connection layer classification, to ensure that the model can accurately classify the comments. Experiments show that TextCNN performs best among all models, with F1 reaching 0.8565. Future research will focus on the sentiment analysis of some complex grammatical structures, and carry out research activities from more dimensions to further improve the accuracy of sentiment analysis model analysis.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Research on How Evaluation Videos Affect the Purchase Intentions of Consumers

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Weidong Huang,
  • Jinyuan Yang,
  • Xinhang Yu,
  • Kanyue Zhang,
  • Qing Zhai

摘要

With the continuous development of Internet technology, evaluation video has become an important marketing tool for merchants and brands in all walks of life, and plays an important role in the purchase intention of consumers. The purpose of this study is to use deep learning technology to conduct a binary task of willing or unwilling to buy for evaluation video comments on Bilibili platform. Firstly, by using a suitable web crawler tool, all the comments are obtained from the target video, and then the necessary pre-processing work is carried out on these comments; Then, a deep learning model with strong expression ability is constructed by using neural network model, and the feature extraction of keywords in comments is carried out by CNN to capture important information in comments. Finally, through the full connection layer classification, to ensure that the model can accurately classify the comments. Experiments show that TextCNN performs best among all models, with F1 reaching 0.8565. Future research will focus on the sentiment analysis of some complex grammatical structures, and carry out research activities from more dimensions to further improve the accuracy of sentiment analysis model analysis.