AgentPEN: A Prediction-Explanation Network for Sequential Stock Movement via LLMs and Recurrent Generation

Summary

AgentPEN is a prediction-explanation network for stock movement that aligns financial news text with price streams using an LLM-based Representation Fusion Agent. It feeds the fused representation into a Deep Recurrent Generation module to predict stock movements. The authors report that it surpasses state-of-the-art baselines in prediction accuracy and explainability on multiple real-world datasets.