IMPLEMENTATION OF THE STEP FUNCTION INTERVENTION AND EXTREME LEARNING MACHINE FOR FORECASTING THE PASSENGER’S AIRPORT IN SORONG

Abstract

This study aims to forecast the number of passengers departing at the domestic departure terminal at Domine Eduard Osok Sorong Airport in 2022 using the Autoregressive Integrated Moving Average (ARIMA) method, ARIMA with Step Function Intervention, and Extreme Learning Machine (ELM). The knowledge of the number of passengers can help the airport prepare facilities. The residual ARIMA model (0,1,0) has no serial correlation (random walk) based on the Ljung-Box test. The MAPE value of the ARIMA model (0,1,0) is 65.47% which means poorly fitted. Because of it, the researchers propose an intervention in the ARIMA model. The RMSE and MAPE ARIMA Intervention ​​(1,0,0) (0,1,0) [12] were 9,027.671 and 35.86%, respectively. Besides, this study also employed the ELM method, which has a MAPE error measurement value of 30.64%. The ELM method has the lowest error measurement results among the three methods. Therefore, the ELM method is suitable for forecasting the number of passengers with predicted values ​​from June to September 2022 as follows: 47985, 37821, 31247, and 33578. On the other hand, intervention in ARIMA can reduce MAPE by 45%.