R. A. Torres Díaz, P. C. Alvarez Esteban
Air transportation growth is a reality described by different sources (e.g. The World Bank report (2018)), then it is advisable to anticipate airport operations to improve safety. Particularly, runway usage is one of these operations.
However, detailed information of runway usage is inaccessible. This is a drawback for a forecast task given that there is no feasible target to fit a model. Thus, the goal of this work is twofold: 1) To introduce an algorithm that allows to reconstruct the target for any airport. 2) To fit models able to predict airport runway usage, based on weather joined to operation data at the airport.
The quality of the results in the first part report a target reconstruction above the 90\% of accuracy for the available comparisons and the forecasting task was successfully applied massively, and we present the results of a standard and a complex case scenarios, Barajas and Schiphol airports.
Palabras clave: Big data, machine learning, airport runway usage, ADS-B technology
Programado
AM-1 Análisis Multivariante
4 de septiembre de 2019 12:00
I3L8. Edificio Georgina Blanes