Ning Xu, Lance Sherry, Kathryn B. Laskey 1 MULTI-FACTOR MODEL FOR PREDICTING DELAYS AT U.S. AIRPORTS Ning Xu Email: nxu@gmu.edu Phone: 703-993-1663 Lance Sherry Email: lsherry@gmu.edu Phone: 703-993-1711 Kathryn Blackmond Laskey Email: klaskey@gmu.edu Phone: 703-993-1644 Fax: 703-993-1521 Center of Air Transportation and Systems Research Department of Systems Engineering and Operations Research George Mason University 4400 University Dr. Fairfax VA 22030 Text: 5290 words (Abstract: 93 words) 1 Tables: 2*250=500 words 4 Figures: 4*250=1000 Total: 5290+500+1000=6790 Abstract: Air transportation is provided by the movement of aircraft through a network of airports. Researchers have established that approximately 84% of delays are generated at airports. These delays propagate to downstream airports where they are absorbed, passed on or enhanced. Researchers have correlated delays with sets of causal factors and have created models to predict aggregate daily delays at airports. Airport Operation Center (AOC) personnel and Traffic Flow Management (TFM) Specialists have suggested that a model for predicting airport delay in 15 minute epochs would have utility. This paper describes multi-factor models for predicting airport delays in 15 minute epochs at each of the 34 OEP airports. The models are developed using piece-wise linear regressions, using Multi-Adaptive Regression Splines (MARS), for generated delays and for absorbed delays for each of the 34 OEP airports. The models were generated using historic individual airport data. Accuracy evaluation on separate test data shows mean absolute prediction error of 5.3 minutes for generated delay across all the airports, and 2.2 minutes for absorbed delay across all the airports. A summary of the factors that drive the performance of each airport is provided. The implications of these results are discussed.