Airport Delay Prediction with Temporal Fusion Transformers

• DOI: 10.48550/arxiv.2405.08293
• Publication Date: 2024-05-13
• Authors: Peng Liu; Kaijing Ding; Xi Cheng; Jianan Chen; Siyuan Feng; Hui Lin; Jilin Song; Chen Zhu
• Keywords: FOS: Computer and information sciences; Computer Science - Machine Learning; Machine Learning (cs.LG)

Since flight delay hurts passengers, airlines, and airports, its prediction becomes crucial for the decision-making of all stakeholders in aviation industry thus has been attempted by various previous research. However, predictions are often categorical at a highly aggregated level. To improve that, this study proposes to apply novel Temporal Fusion Transformer model predict numerical airport arrival delays quarter hour level U.S. top 30 airports. Inputs our include demand capacity forecasts, historic operation efficiency information, wind visibility conditions, as well enroute weather traffic conditions. The results show that achieves satisfactory performance measured small errors on test set. In addition, interpretability analysis outputs identifies important input factors prediction.