Scholarlink

Sentiment analysis plays a crucial role in extracting opinions from textual data, particularly within the healthcare domain. In this study, we focus on aspect-based sentiment analysis (ABSA) of patient drug reviews to identify sentiments related to the effectiveness of medications. Using a multiclass classification approach, we fine-tuned three transformer-based models: DistilBERT, ALBERT, and RoBERTa to classify sentiments into positive, negative, and neutral categories. Experimental evaluation on the Drug Review Dataset sourced from Drugs.com, available on Kaggle, consisting of 214,063 instances, revealed that DistilBERT achieved the best performance with an accuracy of 89%. These results highlight the effectiveness of transformer-based models in capturing nuanced patient feedback, thereby supporting more accurate sentiment interpretation for enhanced clinical decision-making and personalized patient care.

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

Transformer-Based Aspect Sentiment Analysis for Drug Review Mining: A Comparative Evaluation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Jesmin Akter,
  • Sanjida Binta Aziz,
  • Jakia Rahman,
  • Mohammad Aman Ullah

摘要

Sentiment analysis plays a crucial role in extracting opinions from textual data, particularly within the healthcare domain. In this study, we focus on aspect-based sentiment analysis (ABSA) of patient drug reviews to identify sentiments related to the effectiveness of medications. Using a multiclass classification approach, we fine-tuned three transformer-based models: DistilBERT, ALBERT, and RoBERTa to classify sentiments into positive, negative, and neutral categories. Experimental evaluation on the Drug Review Dataset sourced from Drugs.com, available on Kaggle, consisting of 214,063 instances, revealed that DistilBERT achieved the best performance with an accuracy of 89%. These results highlight the effectiveness of transformer-based models in capturing nuanced patient feedback, thereby supporting more accurate sentiment interpretation for enhanced clinical decision-making and personalized patient care.