Exogenous Variable

Exogenous variables, factors external to a system under study, are increasingly central to improving the accuracy and efficiency of various predictive models. Current research focuses on incorporating these variables into diverse frameworks, including machine learning algorithms (e.g., transformers, graph attention networks) and reinforcement learning, often leveraging structured representations like Exo-MDPs to handle the interplay between endogenous and exogenous influences. This work is significant because it enhances forecasting capabilities across numerous domains, from financial markets and traffic prediction to resource management and space situational awareness, by accounting for previously neglected contextual information.