Scholarlink

This paper explores the use of Transformer TF-BERT and PyTorch models to predict Vietnamese students’ sentiments regarding library service quality. Through advanced preprocessing techniques like stemming and tokenization, our experimental results demonstrate that the Transformer-MOD model achieves a superior accuracy of 71%, outperforming traditional models. The study highlights the potential of Transformer-based architectures for large-scale sentiment analysis in educational contexts.

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

Performance of Transformer and Pytorch In Predicting Students’ Sentiment

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Nguyen Minh Tuan,
  • Phayung Meesad,
  • Duong Van Hieu

摘要

This paper explores the use of Transformer TF-BERT and PyTorch models to predict Vietnamese students’ sentiments regarding library service quality. Through advanced preprocessing techniques like stemming and tokenization, our experimental results demonstrate that the Transformer-MOD model achieves a superior accuracy of 71%, outperforming traditional models. The study highlights the potential of Transformer-based architectures for large-scale sentiment analysis in educational contexts.