A Comparison of the ANFIS Model with SARIMA for the Forecasting of Inbound Tourism Demand for Mauritius

Abstract

Tourism planning is important while budgeting potential revenue and expenditure for the economy of a country, especially when large amount of funds are involved in the development of this sector. In this paper, we investigate the application of the adaptive network-based fuzzy inference system (ANFIS) model and the seasonal autoregressive integrated moving average (SARIMA) to predict monthly tourism demand from four continents to Mauritius and compare their performances. Based on the error measures, we find that ANFIS method gives better results than SARIMA method and it can be considered as the preferred tool for forecasting tourism demand for Mauritius.

Keywords: Tourism Demand, Forecasting, ANFIS, SARIMA, Modelling, Fuzzy Systems