{"ID":2890993,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.22932","arxiv_id":"2507.22932","title":"FinMarBa: A Market-Informed Dataset for Financial Sentiment Classification","abstract":"This paper presents a novel hierarchical framework for portfolio optimization, integrating lightweight Large Language Models (LLMs) with Deep Reinforcement Learning (DRL) to combine sentiment signals from financial news with traditional market indicators. Our three-tier architecture employs base RL agents to process hybrid data, meta-agents to aggregate their decisions, and a super-agent to merge decisions based on market data and sentiment analysis. Evaluated on data from 2018 to 2024, after training on 2000-2017, the framework achieves a 26% annualized return and a Sharpe ratio of 1.2, outperforming equal-weighted and S\u0026P 500 benchmarks. Key contributions include scalable cross-modal integration, a hierarchical RL structure for enhanced stability, and open-source reproducibility.","short_abstract":"This paper presents a novel hierarchical framework for portfolio optimization, integrating lightweight Large Language Models (LLMs) with Deep Reinforcement Learning (DRL) to combine sentiment signals from financial news with traditional market indicators. Our three-tier architecture employs base RL agents to process hy...","url_abs":"https://arxiv.org/abs/2507.22932","url_pdf":"https://arxiv.org/pdf/2507.22932v1","authors":"[\"Baptiste Lefort\",\"Eric Benhamou\",\"Beatrice Guez\",\"Jean-Jacques Ohana\",\"Ethan Setrouk\",\"Alban Etienne\"]","published":"2025-07-24T16:27:32Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"q-fin.GN\"]","methods":"[\"Reinforcement Learning\",\"Large Language Model\",\"Language Model\"]","has_code":false}