Stock Prediction via a Dual Relation Fusion Network incorporating Static and Dynamic Relations

2025-10-12

AI Summary

Existing stock prediction methods often neglect the complementary nature of static (e.g., industry, market capitalization) and dynamic (e.g., intraday co-movement) relationships. To address this, we propose the Dual-Relation Fusion Network (DRFN), the first model to jointly and recurrently encode predefined static graph structures with time-varying dynamic relations—constructed from overnight information and distance-aware similarity. DRFN introduces a recursive fusion mechanism that simultaneously captures long-term stable dependencies and short-term market fluctuations. A dynamic weighting scheme adaptively modulates relation strengths, enhancing sensitivity to both structural relevance and price co-movement patterns. Extensive experiments across multiple markets—including China’s A-share and U.S. equity markets—demonstrate that DRFN consistently outperforms state-of-the-art baselines. Results confirm its cross-market effectiveness and robustness, establishing a novel paradigm for integrating heterogeneous, multi-source relational information in financial forecasting.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Machine Learning: Statistical Relational/Logic Learning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Models for Web evolution

Abstract

Accurate modeling of inter-stock relationships is critical for stock price forecasting. However, existing methods predominantly focus on single-state relationships, neglecting the essential complementarity between dynamic and static inter-stock relations. To solve this problem, we propose a Dual Relation Fusion Network (DRFN) to capture the long-term relative stability of stock relation structures while retaining the flexibility to respond to sudden market shifts. Our approach features a novel relative static relation component that models time-varying long-term patterns and incorporates overnight informational influences. We capture dynamic inter-stock relationships through distance-aware mechanisms, while evolving long-term structures via recurrent fusion of dynamic relations from the prior day with the pre-defined static relations. Experiments demonstrate that our method significantly outperforms the baselines across different markets, with high sensitivity to the co-movement of relational strength and stock price.

Problem

• Modeling both dynamic and static inter-stock relationships for prediction
• Capturing long-term stability while responding to sudden market shifts
• Integrating overnight information and time-varying patterns in stock relations

Innovation

• Dual Relation Fusion Network combining static and dynamic relations
• Relative static component models time-varying long-term patterns
• Recurrent fusion of dynamic relations with predefined static relations