In this paper, we present two supervised-learning models, logistic regression and decision tree, to predict occurrence of ground delay program at an airport based on meteorological conditions and scheduled traffic demand. Such predictive capabilities can help the Federal Aviation Administration traffic managers and airline dispatchers to prepare mitigation strategies to reduce the impact of adverse weather. The models are applied to predict ground delay program occurrence at two major U.S. airports: Newark Liberty Intl. and San Francisco Intl. airports. The logistic regression model estimates the probability that a ground delay program will occur during a given hour. Decision tree, on the other hand, classifies an hour as a ground delay program or not based on the input variables. Results indicate that both models perform significantly better than a purely random prediction of ground delay program occurrence at the two airports. The logistic regression model performs better than the decision tree model. The degree to which various input variables impact the probability of ground delay program vary between the two airports. While the enroute convective weather is a dominant factor causing ground delay programs at New York airports, poor visibility and low cloud ceiling caused by marine stratus are major drivers of ground delay programs at San Francisco Intl. airport.
Predicting Ground Delay Program at an Airport Based on Meteorological Conditions
AIAA Aviation 2014 ; 2014 ; Atlanta, GA, United States
2014-06-16
Conference paper
No indication
English
PREDICTING GROUND DELAY PROGRAM AT AN AIRPORT BASED ON METEOROLOGICAL CONDITIONS (AIAA 2014-2713)
British Library Conference Proceedings | 2014
|Ground delay program planning under uncertainty in airport capacity
Online Contents | 2012
|Ground delay program planning under uncertainty in airport capacity
Taylor & Francis Verlag | 2012
|