Celem przedstawionego w artykule badania jest próba analizy przestrzennego zrożnicowania powiatów 10 lat po transformacji. Analizę przeprowadzono na podstawie danych statystycznych obejmujących 307 powiatów badanych ze względu na 9 zmiennych demograficznych i 14 zmiennych ekonomicznych według stanu na rok 2000. Pomiar przestrzennego zróżnicowania wykonano przy wykorzystaniu sieci neuronowej typu SOM (Self-Organizing Map).
EN
The analyse of similarities and diversification of poviats in Poland can serve as basis of diagnosis method of its condition after 10 years transformation period in Poland. A neural networks can be one of the measure used for such analyse. The article presents a trial of application of artificial neural networks for spatial survey of diversification of poviats in Poland. The analyse is conducting on the basis of statistical data concerning poviats, surveyed considering 9 demographic variables and 14 economic variables. The analyse covered Polish poviats without the Warszawski poviat and cities with poviat status - in total 307 administrative units, by situation as of 2000. Classification is based on neural network of Self-Organizing Map type (SOM). As a result it received 3 groups of poviats separated considering demographic variables and 5 groups considering economic variables. Then the analyse of poviats grouping in separate classes was conducted. On the above-mentioned basis it proved that received groups are clearly differed. Having regard to the conducted analyse it is possible to draw some conclusions which can be used for change of regional policy, development of regions and setback of enlarging disproportion between poviats. (original abstract)