Forecasting Inbound Tour Daily Demand with Multi Seasonality Pattern: A Case Study of a Tour Operator in Thailand

Abstract

Tour operators is playing an important role in Tourism industry which is the essential part of industries for Thai economy. Accurate tourist forecasting is very important input for resource planning (e.g., tour guides, vehicle, etc.) for Tour operators. This paper proposes and compares time-series models to forecast daily demand (number of tourists) for a case study tour operator using the Seasonal Autoregressive Integrated Moving Average model (SARIMA), Seasonal Autoregressive Integrated Moving Average model with exogenous variables model (SARIMAX) and Trigonometric ARMA errors, trend and multiple seasonal patterns (TBATS). The performances are evaluated in terms of Mean Absolute Error (MAE) and Mean Absolute Scaled Error (MASE). The results show that TBATS is the overall most accurate model to forecast the number of tourists using this tour operator's services. Comparing with the same day last year method (the present method which is the case-study company's existing model), TBATS can reduce errors by 48.9% for tour A, 30.6% for tour B and 15.8% for tour C, respectively.