APPLICATION OF KOHONEN MAPS FOR THE ANALYSIS OF THE MAIN SOCIO-ECONOMIC INDICATORS OF THE ADMINISTRATIVE DISTRICTS OF THE STAVROPOL TERRITORY
Full recordShow full item record
AbstractThe aim of the work is the clustering of regions of Stavropol Krai on a complex of the main socio-economic indexes. As a tool for cluster analysis using Koho-nen neural network, emulated by the analytical Deductor Studio platform. On the basis of the selected by the authors of the parameters of the neural network training received a partition of the administrative districts of the Stavropol territory on the 4 cluster in the level of their socio-economic development. Leaders Budenovsky, Izobilnensky, Mineralovodsky areas, on the second place: Kochubeevsky, Krasnogvardeysky, Shpakovsky areas lagging: Andropov, Apanasenkovskij, Arzgirsky, Ipatovskaya, Novoaleksandrovsk, Novoselytsya, Trunovskogo, Levokumskoe, Stepnovskogo, Turkmen areas; outsiders Alexandra, Blagodarnenskogo, George, Grachevsky, Kirov, Kursk, Neftekumsk, Petrovsky, Foothill, the Soviet areas. On the basis of visualization of multidimensional data by means of Kokhonen's self-organizing cards, profiles of clusters and the cross-chart the analysis of the received clusters in the analytical Deductor Studio platform is carried out. Objectivity of the results received by authors is confirmed by their identity with results of cluster analysis by means of algorithm of k-means.