Predicting Flight Delays Using Historical Data and Machine Learning Techniques

This study investigates the application of machine learning (ML) techniques to predict flight delays using historical aviation data, with a focus on developing interpretable, scalable models that address operational inefficiencies in the aviation industry. Flight delays incur significant economic costs and disrupt passenger experiences, necessitating robust predictive analytics to enhance decision-making. By leveraging enriched datasets encompassing seasonal trends, carrier-specific performance metrics, and route-level probabilities, this research employs advanced ML algorithms such as gradient boosting machines (GBMs), deep neural networks (DNNs), and hybrid models to identify critical predictors of delays.