The topology of the Russian regions by the expected life expectancy of the population with the use of the Kohonen neural network

The authors proposed a method to analyze the life expectancy of the population of regions of Russia, which is carried out using a Kohonen neural network. By the method of adjusting the input weights of the neural network, the selforganizing maps of Kohonen were chosen. As the determining socio-economic factors affecting the expected life expectancy of the population, statistical indicators are selected taking into account the correlation analysis, which is combined into the following blocks: public health, ecology, social participation and safety, employment and welfare of the population. With the help of Kohonen self-organizing maps, data analysis was performed, clusters were recognized and hidden interdependencies were established between the indicators that affect the expected life expectancy of the population of the regions of the country. Regions of Russia, depending on the level of significance of life expectancy at birth of the population and determining factors of influence on it, were classified into four clusters. Regional economic systems of the South of Russia are classified as a cluster with a high level of significance of life expectancy, due to a low incidence of the population, a balanced level of the employed population and the unemployment rate, as well as a positive ecological situation in the regions.