Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Learning to Generate Explainable Stock Predictions using Self-Reflective Large Language Models

Kelvin J. L. Koa, Yunshan Ma, Ritchie Ng, Tat-Seng Chua

TL;DR

This work proposes a Summarize-Explain-Predict (SEP) framework, which utilizes a verbal self-reflective agent and Proximal Policy Optimization (PPO) that allow a LLM to teach itself how to generate explainable stock predictions, in a fully autonomous manner.

Abstract

Explaining stock predictions is generally a difficult task for traditional non-generative deep learning models, where explanations are limited to visualizing the attention weights on important texts. Today, Large Language Models (LLMs) present a solution to this problem, given their known capabilities to generate human-readable explanations for their decision-making process. However, the task of stock prediction remains challenging for LLMs, as it requires the ability to weigh the varying impacts of chaotic social texts on stock prices. The problem gets progressively harder with the introduction of the explanation component, which requires LLMs to explain verbally why certain factors are more important than the others. On the other hand, to fine-tune LLMs for such a task, one would need expert-annotated samples of explanation for every stock movement in the training set, which is expensive and impractical to scale. To tackle these issues, we propose our Summarize-Explain-Predict (SEP) framework, which utilizes a self-reflective agent and Proximal Policy Optimization (PPO) to let a LLM teach itself how to generate explainable stock predictions in a fully autonomous manner. The reflective agent learns how to explain past stock movements through self-reasoning, while the PPO trainer trains the model to generate the most likely explanations from input texts. The training samples for the PPO trainer are also the responses generated during the reflective process, which eliminates the need for human annotators. Using our SEP framework, we fine-tune a LLM that can outperform both traditional deep-learning and LLM methods in prediction accuracy and Matthews correlation coefficient for the stock classification task. To justify the generalization capability of our framework, we further test it on the portfolio construction task, and demonstrate its effectiveness through various portfolio metrics.

Learning to Generate Explainable Stock Predictions using Self-Reflective Large Language Models

TL;DR

This work proposes a Summarize-Explain-Predict (SEP) framework, which utilizes a verbal self-reflective agent and Proximal Policy Optimization (PPO) that allow a LLM to teach itself how to generate explainable stock predictions, in a fully autonomous manner.

Abstract

Explaining stock predictions is generally a difficult task for traditional non-generative deep learning models, where explanations are limited to visualizing the attention weights on important texts. Today, Large Language Models (LLMs) present a solution to this problem, given their known capabilities to generate human-readable explanations for their decision-making process. However, the task of stock prediction remains challenging for LLMs, as it requires the ability to weigh the varying impacts of chaotic social texts on stock prices. The problem gets progressively harder with the introduction of the explanation component, which requires LLMs to explain verbally why certain factors are more important than the others. On the other hand, to fine-tune LLMs for such a task, one would need expert-annotated samples of explanation for every stock movement in the training set, which is expensive and impractical to scale. To tackle these issues, we propose our Summarize-Explain-Predict (SEP) framework, which utilizes a self-reflective agent and Proximal Policy Optimization (PPO) to let a LLM teach itself how to generate explainable stock predictions in a fully autonomous manner. The reflective agent learns how to explain past stock movements through self-reasoning, while the PPO trainer trains the model to generate the most likely explanations from input texts. The training samples for the PPO trainer are also the responses generated during the reflective process, which eliminates the need for human annotators. Using our SEP framework, we fine-tune a LLM that can outperform both traditional deep-learning and LLM methods in prediction accuracy and Matthews correlation coefficient for the stock classification task. To justify the generalization capability of our framework, we further test it on the portfolio construction task, and demonstrate its effectiveness through various portfolio metrics.
Paper Structure (33 sections, 6 equations, 8 figures, 11 tables)

This paper contains 33 sections, 6 equations, 8 figures, 11 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Methodology
  4. Preliminaries
  5. Problem Formulation
  6. Data Collection and Clustering
  7. Summary Generation
  8. Explanation Generation
  9. Explanation Prompting
  10. Self-Reflective Process
  11. Prediction Generation
  12. Model Fine-Tuning
  13. Confidence-based Sampling
  14. Experiment
  15. Experimental Settings
  16. ...and 18 more sections

Figures (8)

  • Figure 1: While LLMs can classify the sentiment of individual texts (highlighted in orange), they are not trained to weigh between them to produce an aggregate prediction (highlighted in blue). An improved response by our fine-tuned LLM will be presented in the results. [...] refers to truncated text.
  • Figure 2: Overall framework of our proposed SEP method, which consists of three components: Summarize, Explain and Predict.
  • Figure 3: Diagram of the self-reflective process.
  • Figure 4: Diagram of the fine-tuning process.
  • Figure 5: An example of a hallucinated response from Vicuna. [...] refers to truncated text, which are all repeated text.
  • ...and 3 more figures