🤖 AI Summary
Traditional time-series models neglect inter-firm value-chain dependencies, limiting their ability to accurately forecast stock returns. To address this, we propose LSTM-GCN, the first deep learning framework that explicitly incorporates real-world industry topology into stock price prediction—jointly modeling spatial dependencies among firms via Graph Convolutional Networks (GCN) and temporal dynamics of stock prices via Long Short-Term Memory (LSTM) in an end-to-end manner. The model integrates heterogeneous multi-source data, overcoming the limitations of univariate price-only modeling. Evaluated on EuroStoxx 600 and S&P 500 datasets, it significantly outperforms standard baselines—including ARIMA, vanilla LSTM, and GraphSAGE—demonstrating that value-chain structure encodes incremental predictive signals not fully captured by market prices. Our core contribution is the design of the first spatiotemporal neural architecture grounded in empirically derived industry graphs, establishing a novel fundamentals-driven paradigm for quantitative forecasting.
📝 Abstract
Stock price prediction is a crucial element in financial trading as it allows traders to make informed decisions about buying, selling, and holding stocks. Accurate predictions of future stock prices can help traders optimize their trading strategies and maximize their profits. In this paper, we introduce a neural network-based stock return prediction method, the Long Short-Term Memory Graph Convolutional Neural Network (LSTM-GCN) model, which combines the Graph Convolutional Network (GCN) and Long Short-Term Memory (LSTM) Cells. Specifically, the GCN is used to capture complex topological structures and spatial dependence from value chain data, while the LSTM captures temporal dependence and dynamic changes in stock returns data. We evaluated the LSTM-GCN model on two datasets consisting of constituents of Eurostoxx 600 and S&P 500. Our experiments demonstrate that the LSTM-GCN model can capture additional information from value chain data that are not fully reflected in price data, and the predictions outperform baseline models on both datasets.