Di Ciccio, Claudio. 2013. Flight Diversion Prediction Based on Event Data. Eindhoven University of Technology, Eindhoven, The Netherlands, 09.12.

RIS

@MISC{Di Ciccio2013,
title = {Flight Diversion Prediction Based on Event Data},
author = {Claudio Di Ciccio},
year = {2013},
address = {Eindhoven, The Netherlands},
language = {EN},
misc = {Eindhoven University of Technology},
abstract = {An aircraft is diverted when it lands in a place that differs from the expected airport. Diversions can be due to several reasons, such as bad weather conditions, both in the destination airport area or along the route, fuel shortage or other technical problems, etc. In multi-modal transportation scenarios, a cargo plane may be involved in a comprehensive transportation process: e.g., there may be trucks waiting at the destination airport for the inland delivery of goods. Therefore, if the plane were diverted, the logistics companies managing the following transportation legs should be informed as timely as possible. The sooner they are notified, the more chances they have to efficiently rearrange their transportation plans: e.g., rescheduling or rebooking trucks in order to let them reach the actual destination airport. In this presentation, a novel approach for the automated detection of diversions is described. Based on Support Vector Machines, it is designed to rely on publicly available data about aeroplanes' flight information. Though kept unaware of routes, our technique is able to classify a flight as diverted with a very high accuracy and in few minutes after the anomalous behaviour begins. Experiments conducted on real data are reported.},
}