Integrating Large Language Models and Reinforcement Learning for Sentiment-Driven Quantitative Trading
Wo Long, Wenxin Zeng, Xiaoyu Zhang, Ziyao Zhou
TL;DR
The paper tackles the problem of leveraging unstructured sentiment from financial news for trading by deploying FinGPT to generate daily sentiment signals and combining them with traditional technical indicators. It introduces and compares two integration approaches: a conventional rule-based method and a reinforcement-learning–driven framework based on TD3 that dynamically fuses heterogeneous signals. Empirical results show that FinGPT sentiment provides incremental value over purely technical strategies, and the RL-based approach can achieve superior out-of-sample performance despite higher turnover and costs. The work demonstrates the practical potential of signal fusion in quantitative trading and offers guidance on bias mitigation and RL-based decision-making under realistic trading frictions.
Abstract
This research develops a sentiment-driven quantitative trading system that leverages a large language model, FinGPT, for sentiment analysis, and explores a novel method for signal integration using a reinforcement learning algorithm, Twin Delayed Deep Deterministic Policy Gradient (TD3). We compare the performance of strategies that integrate sentiment and technical signals using both a conventional rule-based approach and a reinforcement learning framework. The results suggest that sentiment signals generated by FinGPT offer value when combined with traditional technical indicators, and that reinforcement learning algorithm presents a promising approach for effectively integrating heterogeneous signals in dynamic trading environments.