Scholarlink

Financial sentiment is a key qualitative measure to analyse the opinions and emotions of investors and market participants regarding financial markets and assets. Classifying financial sentiments with linguistic variations provide valuable insights to multi-lingual communities, wherein multiple languages are mainly taken for financial discussions. In this line, this study presents a new approach for financial sentiment classification. We propose a new multi-scale context-aware CoMFinSe-MusCaAt model to classify financial sentiment over English, low-resource (Hindi), and code-mixed (Hindi and English) related three datasets. The performance of the proposed CoMFinSe-MusCaAt model shows impressive results across all datasets and languages. It also outperforms relevant studies and baseline methods in terms of F-score and Accuracy.

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

CoMFinSe-MusCaAt: Code-Mixed Financial Sentiment Classification via Multi-scale Context-Aware Attention on Low-Resource Language Settings

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Ashraf Kamal,
  • Padmapriya Mohankumar,
  • Vishal Kumar Singh

摘要

Financial sentiment is a key qualitative measure to analyse the opinions and emotions of investors and market participants regarding financial markets and assets. Classifying financial sentiments with linguistic variations provide valuable insights to multi-lingual communities, wherein multiple languages are mainly taken for financial discussions. In this line, this study presents a new approach for financial sentiment classification. We propose a new multi-scale context-aware CoMFinSe-MusCaAt model to classify financial sentiment over English, low-resource (Hindi), and code-mixed (Hindi and English) related three datasets. The performance of the proposed CoMFinSe-MusCaAt model shows impressive results across all datasets and languages. It also outperforms relevant studies and baseline methods in terms of F-score and Accuracy.