Scholarlink

The paper is an empirical evaluation of sentiment analysis using text generation models. Metric like, BERTScore is used to gauge semantic similarity. There is accuracy in the generated text by GPT-2, BART, and other models to capture true emotional expressions. BERTScore is an important metric for evaluating the quality of text summarization and similarity of summarized text. There is a comparison between different metrics like BERTScore, BLUE, and ROUGE.

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

Benchmarking Text Generation Models for Sentiment Analysis: A BERTScore-Centric Approach

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Megha Kesar,
  • Vivek Kumar,
  • Hitesh Singh,
  • Megha Gupta

摘要

The paper is an empirical evaluation of sentiment analysis using text generation models. Metric like, BERTScore is used to gauge semantic similarity. There is accuracy in the generated text by GPT-2, BART, and other models to capture true emotional expressions. BERTScore is an important metric for evaluating the quality of text summarization and similarity of summarized text. There is a comparison between different metrics like BERTScore, BLUE, and ROUGE.