Optimizing Performance: How Compact Models Match or Exceed GPT's Classification Capabilities through Fine-Tuning

TL;DR

This study shows that small, non-generative models like FinBERT, when fine-tuned, can match or surpass GPT-3.5 and GPT-4 in financial sentiment classification, offering a more efficient alternative.

Contribution

The paper introduces a novel financial news database and demonstrates that fine-tuned compact models can outperform large GPT models in zero-shot sentiment analysis tasks.

Findings

Fine-tuned FinBERT outperforms GPT-3.5 in financial sentiment analysis.

A new database for financial news sentiment scoring was created.

Fine-tuned small models show behavioral similarities to GPT models.

Abstract

In this paper, we demonstrate that non-generative, small-sized models such as FinBERT and FinDRoBERTa, when fine-tuned, can outperform GPT-3.5 and GPT-4 models in zero-shot learning settings in sentiment analysis for financial news. These fine-tuned models show comparable results to GPT-3.5 when it is fine-tuned on the task of determining market sentiment from daily financial news summaries sourced from Bloomberg. To fine-tune and compare these models, we created a novel database, which assigns a market score to each piece of news without human interpretation bias, systematically identifying the mentioned companies and analyzing whether their stocks have gone up, down, or remained neutral. Furthermore, the paper shows that the assumptions of Condorcet's Jury Theorem do not hold suggesting that fine-tuned small models are not independent of the fine-tuned GPT models, indicating behavioural similarities. Lastly, the resulted fine-tuned models are made publicly available on HuggingFace, providing a resource for further research in financial sentiment analysis and text classification.