Ploutos: Towards interpretable stock movement prediction with financial large language model
Hanshuang Tong, Jun Li, Ning Wu, Ming Gong, Dongmei Zhang, Qi Zhang
TL;DR
The document provides standardized formatting guidelines for IJCAI-24 submissions, outlining requirements for length ($7$ content pages plus up to $2$ for references and ancillary sections), anonymity during review, and the camera-ready version. It prescribes a comprehensive template for both LaTeX and Word, a strict two-column layout, font and spacing rules, and detailed instructions for sections, citations, figures, tables, formulas, and special sections. The contribution ensures consistency, fairness in the review process, and efficient production of the proceedings by defining precise ordering, formatting, and template resources. Collectively, these guidelines facilitate high-quality, uniform presentation of research while accommodating track-specific variations and optional sections such as Appendices, Ethical Statements, and Acknowledgements.
Abstract
Recent advancements in large language models (LLMs) have opened new pathways for many domains. However, the full potential of LLMs in financial investments remains largely untapped. There are two main challenges for typical deep learning-based methods for quantitative finance. First, they struggle to fuse textual and numerical information flexibly for stock movement prediction. Second, traditional methods lack clarity and interpretability, which impedes their application in scenarios where the justification for predictions is essential. To solve the above challenges, we propose Ploutos, a novel financial LLM framework that consists of PloutosGen and PloutosGPT. The PloutosGen contains multiple primary experts that can analyze different modal data, such as text and numbers, and provide quantitative strategies from different perspectives. Then PloutosGPT combines their insights and predictions and generates interpretable rationales. To generate accurate and faithful rationales, the training strategy of PloutosGPT leverage rearview-mirror prompting mechanism to guide GPT-4 to generate rationales, and a dynamic token weighting mechanism to finetune LLM by increasing key tokens weight. Extensive experiments show our framework outperforms the state-of-the-art methods on both prediction accuracy and interpretability.