# Fine-Tuning Gemma-7B for Enhanced Sentiment Analysis of Financial News Headlines

**Authors:** Kangtong Mo, Wenyan Liu, Xuanzhen Xu, Chang Yu, Yuelin Zou, Fangqing Xia  
**arXiv:** [2406.13626](https://arxiv.org/abs/2406.13626)  
**Date:** 2024-06-19  
**Summary page:** https://www.summarizepaper.com/en/paper/2406.13626v1/

## Key points

- Study focused on sentiment analysis in financial news headlines using NLP and LLMs
- Utilized FinancialPhraseBank dataset for sentiment categorization
- Fine-tuned models including distilbert-base-uncased, Llama, and gemma-7b
- Gemma-7b model outperformed others with highest precision, recall, and F1 score post fine-tuning
- Significant improvement in accuracy of gemma-7b model in capturing financial sentiment nuances
- Refined model can provide valuable market insights, aid in risk management strategies, and facilitate informed investment decisions
- Implications highlight the role of advanced LLMs in reshaping financial information analysis

## Summary

In the study titled "Fine-Tuning Gemma-7B for Enhanced Sentiment Analysis of Financial News Headlines," authors Kangtong Mo, Wenyan Liu, Xuanzhen Xu, Chang Yu, Yuelin Zou, and Fangqing Xia delve into the realm of sentiment analysis within financial news headlines to gain insights into investor sentiment. Utilizing NLP and LLMs, the researchers focus on understanding sentiment from the perspective of retail investors. Their analysis is grounded in the FinancialPhraseBank dataset, which offers categorized sentiments of financial news headlines as a foundation for their research. The team fine-tuned multiple models such as distilbert-base-uncased, Llama, and gemma-7b to assess their efficacy in sentiment classification. Through a series of experiments, they found that the fine-tuned gemma-7b model outperformed others by achieving the highest precision, recall, and F1 score. Particularly noteworthy was the significant improvement in accuracy displayed by the gemma-7b model post fine-tuning, showcasing its robustness in capturing the nuances of financial sentiment. The implications of this refined model are substantial as it can play a pivotal role in providing valuable market insights, aiding in risk management strategies, and facilitating informed investment decisions by accurately predicting the sentiment embedded within financial news. The results underscore  of advanced LLMs in reshaping how we analyze and interpret financial information. Ultimately, for stakeholders within  for enhanced decision-making processes.