Scholarlink

With the rapid growth of social media comments, Sentiment Analysis (SA), as a part of Natural Language Processing (NLP), faces an urgent need for efficient analysis and extraction of vast amounts of internet data. This study proposes a sentiment classification model based on a hybrid BERT model and a hypergraph attention network (HB_HAGT), aiming to accurately identify and classify textual sentiments by combining BERT's semantic representation capabilities with the local attention mechanism of the hypergraph attention network. By constructing a hypergraph model to enrich the feature representation of the data, the model demonstrates its superiority in sentiment analysis tasks through comparative experiments and ablation studies across multiple datasets.

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 Sentiment Classification Based on Hybrid BERT Model and Hypergraph Attention Network

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Wei Dai,
  • Dequan Zheng,
  • Feng Yu,
  • Feng Yan

摘要

With the rapid growth of social media comments, Sentiment Analysis (SA), as a part of Natural Language Processing (NLP), faces an urgent need for efficient analysis and extraction of vast amounts of internet data. This study proposes a sentiment classification model based on a hybrid BERT model and a hypergraph attention network (HB_HAGT), aiming to accurately identify and classify textual sentiments by combining BERT's semantic representation capabilities with the local attention mechanism of the hypergraph attention network. By constructing a hypergraph model to enrich the feature representation of the data, the model demonstrates its superiority in sentiment analysis tasks through comparative experiments and ablation studies across multiple datasets.