Number of Foreign Tourist Arrival Forecasting Using Percentile Error Bootstrap Based on VARIMA Model

Abstract

Forecasting number of foreign tourist arrivals is important to improve the policies in the tourism sector. Better accuracy of forecast would help the government and investor to make operational, tactical, and strategic decisions. Data used in this research are monthly number of foreign tourist arrivals taken from Indonesia Central Bureau of Statistics. Multivariate forecasting at Soekarno-Hatta, Juanda, and Adi Sumarmo arrival gates was conducted using VARIMA ([12],1,0) (0,1,0)12 model. However, the longer step ahead to forecast, the larger variance error of corresponding models. As a result, the prediction interval become wider. This research computed the prediction interval using percentile error bootstrap based on VARIMA models that produced more precise forecast.