Logo image
Back
Airlines Traffic Forecasting Using Deterministic and Stochastic Time Series Decomposition
Journal article   Peer reviewed

Airlines Traffic Forecasting Using Deterministic and Stochastic Time Series Decomposition

William L. Huth and Steven E. Eriksen
The Logistics and Transportation Review, Vol.23(4), pp.401-409
12/1987

Metrics

15 Record Views

Abstract

Accurate forecasts of passenger and cargo loads in the commercial airline industry are crucial. The methodologies available for forecasting air traffic are as diverse as the applications. A time-series procedure is described for forecasting short-term, point-to-point (city-to-city) scheduled air traffic. A novel forecasting approach is suggested that combines decomposition and stochastic modeling methods in a multistage procedure. First is a decomposition during which 2 instruments are created from the original time series. Then, each instrument is modeled individually using stochastic techniques. Finally, the models are recombined to produce a forecast. The technique is applied to quarterly passenger data taken from Table 8 of the Civil Aeronautics Board origin-destination survey. The city pair is New York City-Los Angeles. An improvement in ex post predictive precision is noted. A major feature of the method is its handling of seasonality.

Details

Logo image