A Distillation-based Future-aware Graph Neural Network for Stock Trend Prediction

2025-02-15

AI Summary

Stock trend forecasting is hindered by conventional graph neural networks (GNNs), which model only historical spatiotemporal dependencies while neglecting distributional relationships between past and future patterns. To address this, we propose DishFT-GNN, a future-aware distillation framework that introduces future distribution awareness into GNN training for the first time. Leveraging teacher–student collaborative knowledge distillation, it employs distributional shifts—identified by the teacher model between historical and future patterns—as an intermediate supervisory signal to guide the student in learning future-aware spatiotemporal embeddings. The framework integrates multi-step temporal graph construction with a two-stage iterative training paradigm to implicitly capture dynamic trend evolution. Evaluated on two real-world financial datasets, DishFT-GNN significantly improves accuracy and robustness in upward/downward trend prediction, achieving state-of-the-art performance.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Stock trend prediction involves forecasting the future price movements by analyzing historical data and various market indicators. With the advancement of machine learning, graph neural networks (GNNs) have been extensively employed in stock prediction due to their powerful capability to capture spatiotemporal dependencies of stocks. However, despite the efforts of various GNN stock predictors to enhance predictive performance, the improvements remain limited, as they focus solely on analyzing historical spatiotemporal dependencies, overlooking the correlation between historical and future patterns. In this study, we propose a novel distillation-based future-aware GNN framework (DishFT-GNN) for stock trend prediction. Specifically, DishFT-GNN trains a teacher model and a student model, iteratively. The teacher model learns to capture the correlation between distribution shifts of historical and future data, which is then utilized as intermediate supervision to guide the student model to learn future-aware spatiotemporal embeddings for accurate prediction. Through extensive experiments on two real-world datasets, we verify the state-of-the-art performance of DishFT-GNN.

Problem

• Predicts stock trends using historical data
• Improves GNNs with future data correlation
• Enhances prediction with distillation-based framework

Innovation

• Distillation-based GNN framework
• Future-aware spatiotemporal embeddings
• Teacher-student model training