MDGNN: Multi-Relational Dynamic Graph Neural Network for Comprehensive and Dynamic Stock Investment Prediction
Hao Qian, Hongting Zhou, Qian Zhao, Hao Chen, Hongxiang Yao, Jingwei Wang, Ziqi Liu, Fei Yu, Zhiqiang Zhang, Jun Zhou
TL;DR
The paper addresses stock price movement prediction under dynamic, multifaceted relations by introducing MDGNN, a hierarchical multi-relational dynamic graph model. It constructs daily multi-relational graphs over stocks, industries, and investment banks, encodes them with a multi-path relational GNN, and uses a Transformer-based temporal extractor to model day-to-day evolution; a prediction layer yields the probability of positive returns. Key contributions include a novel intra-day multi-relational graph embedding, an inter-day temporal extraction layer with ALIBI bias, and extensive experiments showing state-of-the-art performance on CSI100/CSI300 benchmarks, along with ablations and case studies. The approach enables comprehensive representation of inter-stock relations and their temporal dynamics, with practical impact for investment prediction and risk assessment.
Abstract
The stock market is a crucial component of the financial system, but predicting the movement of stock prices is challenging due to the dynamic and intricate relations arising from various aspects such as economic indicators, financial reports, global news, and investor sentiment. Traditional sequential methods and graph-based models have been applied in stock movement prediction, but they have limitations in capturing the multifaceted and temporal influences in stock price movements. To address these challenges, the Multi-relational Dynamic Graph Neural Network (MDGNN) framework is proposed, which utilizes a discrete dynamic graph to comprehensively capture multifaceted relations among stocks and their evolution over time. The representation generated from the graph offers a complete perspective on the interrelationships among stocks and associated entities. Additionally, the power of the Transformer structure is leveraged to encode the temporal evolution of multiplex relations, providing a dynamic and effective approach to predicting stock investment. Further, our proposed MDGNN framework achieves the best performance in public datasets compared with state-of-the-art (SOTA) stock investment methods.