Anomaly Detection in Global Financial Markets with Graph Neural Networks and Nonextensive Entropy
This work addresses anomaly detection for financial market analysts, but it is incremental as it combines existing methods (GNNs and nonextensive entropy) on financial data.
The study tackled anomaly detection in global financial markets by using Graph Neural Networks and nonextensive entropy, finding that asset correlation structures change during crises and anomaly counts vary statistically with entropy parameters across crisis phases.
Anomaly detection is a challenging task, particularly in systems with many variables. Anomalies are outliers that statistically differ from the analyzed data and can arise from rare events, malfunctions, or system misuse. This study investigated the ability to detect anomalies in global financial markets through Graph Neural Networks (GNN) considering an uncertainty scenario measured by a nonextensive entropy. The main findings show that the complex structure of highly correlated assets decreases in a crisis, and the number of anomalies is statistically different for nonextensive entropy parameters considering before, during, and after crisis.