An Analysis of Different Sentiment Analysis Models on Financial Text using Transformer

DOI

10.2991/978-94-6463-852-3_6How to use a DOI?

Keywords

Financial Sentiment Analysis (FSA); FinBERT; DistilBERT; VADER; and LM

Abstract

In recent years, sentiment analysis has emerged as a crucial instrument within the financial industry, allowing stakeholders to gauge market sentiments and make well-informed choices. Nonetheless, accurately capturing sentiment from financial texts poses significant challenges due to specialized terminology and the complexity of context. This research tackles this issue by comparing four sentiment analysis models—FinBERT, DistilBERT, VADER, and a general Language Model (LM)—across four distinct financial datasets: SEntFiN, FIQA & PhraseBank, Headlines, and Microblogs.

The evaluation of these models was conducted using standard metrics such as accuracy and weighted F1-score. FinBERT demonstrated superior performance, achieving an F1-score of 93.27% and an accuracy of 91.08% on the SEntFiN dataset. DistilBERT, although less demanding in terms of computational resources, recorded the highest accuracy of 93.23% on the same dataset. In contrast, both VADER and the general LM faced difficulties in interpreting nuanced sentiments, especially in contextually rich financial texts.

These results highlight the importance of domain-specific pretraining and transformer-based models in the realm of financial sentiment analysis. The study provides valuable insights for selecting suitable models for practical financial applications and suggests areas for future research and development.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Shriram Bansal
AU  - Bhupesh Kumar Singh
AU  - Mayank Kumar Jain
PY  - 2025
DA  - 2025/10/07
TI  - An Analysis of Different Sentiment Analysis Models on Financial Text using Transformer
BT  - Proceedings of the MULTINOVA: First International Conference on Artificial Intelligence in Engineering, Healthcare and Sciences (ICAIEHS- 2025)
PB  - Atlantis Press
SP  - 83
EP  - 99
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-852-3_6
DO  - 10.2991/978-94-6463-852-3_6
ID  - Bansal2025
ER  -