Prediction of all-causes of airline delays at Zurich airport based on decision tree

Abstract

This paper explores the main factors which may potentially cause flight delay at a hub airport and proposes a generalized flight delay prediction model based on the decision tree (DT) method. The proposed model is developed using data recorded at Zurich airport. From the obtained DT models, we can conclude that the primary splitters in the classification of delayed flights are: the period of the day, whether the flight belongs to the dominant airline, type of handling agent, peak or off-peak period, whether the flight is from North America/Europe or a Schengen country, number of passengers and number of connecting passengers.