Predicting systemic risk in financial systems using Deep Graph Learning

Vicente Balmaseda, María Coronado, Gonzalo de Cadenas-Santiago

Abstract: Highlights•GNNs leverage the network structure to surpass traditional machine learning.•Experiments showing GNNs improve large-scale systemic risk prediction performance.•Estimating a labeled financial network from aggregated data for experiments.•C2R approach to predict a regression output from ordered class labels.•Experiments showing improved systemic risk predictions compared to classification.