Rare events detection and forecasting in dynamic systems

Abstract

Rare events, such as financial crises, banking failures and epidemics outbreaks can significantly impact economies and people’s lives. Early detection in the banking sector is crucial for mitigating financial losses and maintaining operational stability. Traditional models often fail to predict these events due to their unpredictable nature and complex dynamics. Existing systems based on methods like decision trees and neural networks struggle with scalability and accuracy, limited by specific resources of information systems and real datasets.