Journal of Korean Society of Transportation ISSN:1229-1366(Print) 2234-4217(Online) 대한교통학회지

The era of big data in the aviation market is coming, and various methods for forecasting air traffic passenger demand along with many data are being suggested. In addition, short and mid-term demand forecasting studies have increased significantly in recent years, whereas existing air traffic passenger demand forecasting studies have been studied in the long-term view. Currently, demand for international air traffic passengers is skyrocketing every year. But air traffic passenger demand in certain countries has fluctuated greatly due to external influences such as national relations. Existing short and mid-term international air traffic passengers demand forecast is mainly used by ARIMA model and Exponential Smoothing method. In the condition that demand data is influenced externally, these models produce inaccurate forecasting result as they can’t consider about the external influences. In this study, an improved model is proposed for short and mid-term demand forecast of international flights at domestic airports. This paper introduces ARIMA-Intervention that is improved from univariate time series ARIMA model with the consideration of intervention effect and presents the results of an empirical predictive analysis in respect to the international air traffic passenger demand of the Bottom-up method. In addition, we proposed the process for model selection. Monthly data were used from January 2013 to October 2017 for analysis and this paper presents international air traffic demand forecasting result of domestic airport by combining the forecasting results of air traffic passenger demand on China, Japan, the United States, Hong Kong, Thailand, the Philippines, Vietnam and Taiwan with high demand of international flights and in the other counties. At the conclusion, the prediction results of the model using ARIMA-Intervention are presented with MAPE. Based on the proposed model, it was found that the demand prediction for first half of 2018 resulted an error about 1.78% contrasted with the actual international passenger demand, thus, it is able to confirm that it is highly reliable. Finally, we project the 2019 forecast (Business As Usual, BAU) to increase by about 9.34%.