ARIMA 모델을 활용한 인천국제공항 환승여객 수요예측

Abstract

The International Air Transport Association predicts that the number of air travelers in 2024 will reach 4 billion, an increase of 3% compared to before COVID-19((IATA, 2019). This trend is rapidly recovering to the pre-crisis situation in Europe and the Americas with the easing of quarantine measures, and in Korea, the number of passengers using Incheon International Airport per day exceeds 20,000. However, Incheon Airport, which opened in 2001, continues to grow with the goal of becoming a Northeast Asian hub airport, and the number of air passengers is steadily increasing, but the airport's hub function is weakening due to a decrease in the number of transfer passengers or congestion(IATA, 2022). The number of transfer passengers at Incheon International Airport used to recover soon after a temporary decrease due to global diseases or economic recession, but the recent decrease in the number of transfer passengers has led China and Japan to actively revise their aviation policies and to revitalize their domestic airports. It can be said that its nature is different because it can be inter pretedas a phenomenon that occurred in the process of accelerating the strengthening of competitiveness to grow into a North east Asian hub airport. Moreover, as national airlines, which transport most of the transfer passengers at Incheon International Airport, have recently reduced low-profit routes and concentrated on high-profit routes, the supply of seats for transfers and route connectivity have been relatively weakened, while China, the Middle East, and Japan have Airlines are actively expanding direct routes with relatively low fares and visa-free benefits, absorbing a significant portion of transfer passengers at Incheon International Airport, so there are concerns that this reduction in the number of transfer passengers is likely to be prolonged in the future(Homsombat, Lei, & Fu, 2014). This study aims to predict future demand by using the ARIMA model, a time series technique, for the demand of transfer passengers, which is the factor of hubization of airports. Time series forecasting using historical data is very important today. ARIMA models are used in many fields such as finance, industry, medicine, and meteorology(Manjula and Girija, 2022). For example, the ARIMA model is suitable for short-term prediction and is widely used in infectious disease prediction because it can capture periodicity, trend, and randomness of data with high predictive accuracy(Pai, Juan & Deguang, 2022). ARIMA models are also used for time series analysis and prediction of future points in a series or to better understand data. Some of the advantages of ARIMA models are that they use an online learning environment, sample size is independent of storage cost, and parameter estimation can be performed online in a scalable and efficient manner. The down-side is that ARIMA models have a subjective process. The reliability of the selected model may depend on the skill and experience of the predictor, and also depends on several restrictions on the parameters and the class of possible models(Reinsel & Ljung, 2015). ARIMA models have been used as test guides in some modeling methods. We also developed a detailed ARIMA model in the form of a case study using macroeconomic indicators to account for the USD/EUR exchange rate(Weisanget, 2015). There are many methods for forecasting time series data for future periods, but if the time series data is unusual, it is not suitable for forecasting(Romi, 2019). In this study, future demand forecasting was carried out after securing normality through difference of abnormal parts due to trend and seasonality of time series data. The Incheon International Airport transfer demand time series data is monthly time series data from January 2010 to October 2022, and 154 observation samples were used, and the total value of arrival and departure of transfer passengers at Incheon International Airport was used. The data used for the analysis were aviation statistical data from Incheon International Airport Corporation. I think future transfer passenger demand forecasting will be helpful in determining the target and scope of efficient airport operation and niche marketing. Demand forecasting for passengers in the air transportation market is a very important field in planning and operating the market. Demand forecasting methodology should be presented. Through time series analysis, transfer passengers are predicted from 2022 to 2025, and it is expected to be used as objective basic data when establishing the government's aviation policy, and to help establish management strategies for tourism-related businesses.

Key words: demand forecasting, ARIMA Model, Hub Airport, Incheon International Airport, Transit passenger