Modelling and forecasting of Tunisia's health expenditures using artificial neural network and ARDL models

Abstract

This paper brings into play artificial neural network (ANN) to forecast yearly Tunisian health expenditures. We also compare the prediction accuracy of ANN with that of autoregressive distributed lag (ARDL) model. deals with the modelling of the health expenditures in Tunisia in order to forecast future projections based on socio-economic and demographic variables (gross domestic product-GDP, population ageing , medical density and environmental quality) using artificial neural network (ANN) and the . Thus, the future health expenditures of Tunisia are calculated by means of this model under a scenario. Mean Absolute Percent Error (MAPE) and Root Mean Square Error (RMSE) are used in the comparison of both models. The model that has this low statistic is superior to the other model. In the context, the results obtained revealed an ANN model superior to an ARDL model in health expenditures forecasting. Moreover, the scenario used showed that the future health expenditures of Tunisia would increase in a non proportional way with the GDP. Health would be a luxury good in 2020.