Standard of living in Poland at regional level - classification with Kohonen self-organizing maps

The standard of living is spatially diversified and its analyzes enable shaping regional policy. Therefore, it is crucial to assess the standard of living and to classify regions due to their standard of living, based on a wide set of determinants. The most common research methods are those based on composite indicators, however, they are not ideal. Among the current critiques moved to the use of composite indicators is the normative nature of the weights, the drawbacks of the linear method of aggregation and the fact that composite indicators flatten alternatives and differences among analyzed objects. That is why more objective alternative solutions are nowadays gaining popularity. The aim of this article is the regional classification of Polish NUTS-4 regions due to their standard of living using SOM - self-organizing maps - the model of a neural network mapping multidimensional space into a two-dimensional map of neurons. In this study, the set of selected 45 variables describing the standard of living in Polish NUTS-4 regions in 2016 was used.