Unveiling Taiwan Stock Market Dynamics: A Graph Neural Network Approach to Automatic Stock Clustering for Enhanced Predictions

In this paper, we introduce graph neural networks (GNNs) and automatic stock clustering to improve stock market prediction accuracy in the market of Taiwan Stock Exchange. GNNs capture intricate inter-stock relationships, enhancing prediction accuracy. Automatic clustering based on market behavior ensures adaptability. Results demonstrate significant improvements, with potential applicability beyond Taiwan's market, advancing financial prediction methodologies.