THE APPLICATION OF ARTIFICIAL NEURAL NETWORKS IN MODELLING DOMESTIC TOURISM DEMAND

The paper investigates the efficiency of artificial neural networks in modelling domestic tourist demand in the Republic of Croatia, approximated by the number of arrivals and the number of realized overnight stays of domestic tourists. Industrial production volume indices, consumer price indices, average net monthly wages, the number of employees and the monthly seasonal dummy variables were selected as input variables. Two multilayer perceptrons neural networks models were used to model empirical data. The model predictive efficiency was evaluated using the mean average, mean absolute percentage, mean squared root forecast errors, as well as the coefficient of determination and the Pearsons correlation coefficient. The obtained results evaluation showed that the selected multilayer perceptrons models are reliable for modelling domestic tourism demand, although the research is based on a limited small amount of data as well as the number of input variables. Given the research results, as well as the research limitations, it can be concluded that the artificial neural networks models have significant application potentials in modelling time-series of arrivals and overnight stays of domestic tourists in the Republic of Croatia.