A comparative study on modelling and forecasting tourism revenues: The case of Turkey

Abstract

Tourism revenues have important implications for tourism countries in terms of management of tourism-related policies. In order to accurately direct production planning, pricing, promotion and strategic marketing programs, labor and capital resources, accurate and reliable forecasts are needed. Forecasting the developments in tourism with scientific basis methods is an important guide for central and local public administration programs and tourism operators. When reviewing the literature, comparative studies on modeling and forecasting tourism revenues using Artificial Neural Networks (ANNs) are limited and this paper aims to fill this gap. Based on the gap seen in the literature, the purpose of this study is to develop the optimal forecasting model that yields the highest accuracy when comparing the performances of three different methods namely Exponential Smoothing, Box-Jenkins and ANNs for forecasting Turkey's tourism revenues. Forecasting performances of the models were measured by MAPE statistics. As a result of the analyses performed, it was found that ANN Model with [4:5:1] architecture was the best one among the all models applied in this study.