CMStatistics 2022: Start Registration
View Submission - CFE
A1005
Title: Time series forecasting of tourist arrival in Singapore Authors:  Khay Boon Tan - Singapore Institute of Management (Singapore) [presenting]
Abstract: The aim is to apply various forecasting time series methods to forecast tourist arrival in Singapore. Deterministic time trend models, the Smoothing method and ARIMA modelling are used, and a one-step-ahead forecast is performed to compute the mean square forecast error. It is discovered that the log-linear model has the best goodness of fit among the deterministic time trend models, and the Holt and Winters smoothing method incorporating seasonality has the best performance among the class of smoothing methods. For ARIMA modeling, the best performance models are AR(2) model and ARIMA (1,1,1) models. When all the models are applied to generate forecasts, it is discovered that Holt and Winters exponential smoothing incorporating additive seasonality put up the best performance in the forecast of tourist arrival in Singapore.