Large Language Models (LLMs) for Financial Sentiment Analysis and Market Forecasting

DOI

10.2991/978-94-6463-872-1_43How to use a DOI?

Keywords

Financial Sentiment Analysis; Market Forecasting; LLMs; Llama3; Gemma2; RoBERTa; QLoRA

Abstract

Financial markets are always volatile and do not depend on a single macro-economic indicator or economic data but on a whole range of factors, primarily events related to government, geopolitical issues, or corporate earnings, as well as investor sentiment. Unlike traditional quantitative models like time series and econometric models, financial text data is very complex, and most of these models are not suitable for capturing such complexities. Large Language Mod- els (LLMs) are a game changer that has brought forth the effective use of NLP techniques to extract meaning from financial news, earnings reports, and social media sentiments. In this study, we test the efficacy of LLMs (Llama3, Gemma2, RoBERTa) in performing financial segment evaluation and financial market prediction. To improve sentiment classification accuracy, the structured methodology employed consists of data preprocessing, tokenization, and fine- tuning using QLoRA. According to experimental results, Llama3 performs better than other models, achieving an accuracy of 86.1%, RoBERTa attains 85.9%, and Gemma2 gets 84.4%. The study demonstrates that LLMs can effectively capture the sentiment-based market movements however, the challenge lies in the interpretability of the model, removal of the model bias, and performing inference in real time. The study findings further indicate that LLMs look very promising to use in financial forecasting, outperforming conventional models on predicting financial outcomes.

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  - Vishnu Ravi
AU  - Vineet Kumar Srivastava
AU  - Maninder Pal Singh
AU  - Srinivas Chippagiri
AU  - Nikhil Kassetty
AU  - Padma Naresh Vardhineedi
AU  - Ravi Kumar Burila
AU  - Nuzhat Noor Islam Prova
PY  - 2025
DA  - 2025/11/04
TI  - Large Language Models (LLMs) for Financial Sentiment Analysis and Market Forecasting
BT  - Proceedings of the 2nd International Conference on Sustainable Business Practices and Innovative Models (ICSBPIM-2025)
PB  - Atlantis Press
SP  - 681
EP  - 694
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6463-872-1_43
DO  - 10.2991/978-94-6463-872-1_43
ID  - Ravi2025
ER  -