Year
1997
Abstract
This paper examines the monthly predictions of aircraft traffic between Paris and Nantes. This time series is contaminated with outliers due to the influence of the arrival of the atlantic high speed train (call TGV in French). Standard ARIMA modelling techniques, intervention analysis and neural networks are used. To detect some non linearity of time serie, we perform Keenan, Tsay, McLeod-Li, Lee, White-Granger and RESET tests. We eventually compare the forecast from different models.
INDJEHAGOPIAN, J.P. et MACE, S. (1997). Modélisation du trafic aérien en utilisant une approche linéaire et une approche par réseaux neuronaux.