Predicting Flight Delays Using Graph Neural Networks

Created using ChatSlide

This coursework explores the complexities of predicting flight delays using advanced technology-driven approaches. It begins with an introduction to the interconnected nature of airport networks and the limitations of current models. The focus then shifts to Graph Neural Networks (GNNs), examining BasicGNN, SemanticGNN, and GraphSAGE within the U.S. airport system, leveraging both static and dynamic features. A novel "Horizon-Aware" evaluation methodology is applied, using historical data to...