Forecasting Tourist Arrivals from Asian Countries to Sri Lanka

Abstract

Sri Lanka is considered as a wonderful tourist destination in the world. The Asian region is the leading tourist producer to Sri Lanka for many years. To gain maximum benefits and minimize the risk, forecasting of arrivals will be essential. Therefore, the current study focused on fitting a suitable model to forecast tourist arrivals from Asian countries to Sri Lanka. Monthly tourist arrival data from India, Maldives, China, and Japan for the period from, January 2008 to December 2014 was obtained from Sri Lanka Tourism Development Authority (SLTDA). Holt’s Winters Three-Parameter Model was tested on forecasting arrivals from India, Maldives, and Japan. Four trend model; including three linear and one non-linear model were tested to forecast arrivals from China. Model selection criteria were Mean Absolute Percentage Error (MAPE), Mean Absolute Deviation (MAD) and Mean Square Error (MSE). Residual plots and Anderson-Darling test were used for model validation. The results of the study show that Holt's Winters Three-Parameter additive model for forecasting arrivals from India, Maldives, and Japan has MAPE of 1.35%, 1.49%, and 4.6% respectively. Winter's multiplicative model has MAPE 1.32%, 1.36%, and 2.76% respectively for the above countries. The Quadratic trend model is the most suitable model for forecasting arrivals from China. The residuals of all selected models were normally distributed and independent. It was concluded that Winter’s models and Quadratic trend models are suitable to forecast arrivals from India, Maldives China, and Japan. It is recommended to test Seasonal Auto Regressive Integrated Moving Average (SARIMA) and Circular model (C M) for better forecasting.