Abstract:

Forecasting the values of the main macroeconomic indicators has for years been one of the key challenges faced by public authorities, which are aimed at increasing the productivity of economies and long-term sustainable development. There are a number of different categories of factors that influence the level of macroeconomic productivity. Some of them are demographic in nature, others are related to the level of economic activity of labor resources, level of so called economic freedom and the pace of technological development. The purpose of this article is to analyze and assess the influence of selected determinants on GDP per capita value in chosen European countries. The analysis takes into account a number of factors that affect macroeconomic productivity. Achieving the stated purpose was possible thanks to the use of artificial neural networks. Verification of the test results was carried out using Hellwig’s integral capacity method. The time range of the studies covered the years 2004-2015. An important value of this study is also presenting the possibility of forecasting the GDP per capita value using artificial neural networks. This tool can provide a valid approximation to the generating mechanism of a vast class of non-linear processes, for their use as forecasting devices. In case of research problem from this article – a multi-layer perceptron (MLP) seemed to be the best choice for modeling. It is worth emphasizing that that the highest level of correlation with GDP per capita was recorded for the factor related to R&D expenditure, which was the most influential predictor in the model. The obtained results seem to confirm the key importance of the level of R&D spending in designing and implementing the European policy of dynamic growth and sustainable development.