Living standards control in the region: modeling of factors of external and internal environment

This paper proposes theoretical and methodical approaches to forecasting the living standards on the basis of neural networks. The assessment of the main indices of living standards is a subject of study of both federal statistical agencies and different expert organizations. However, an adequate measurement of these indicators is confronted with information and methodological problems, including: the incompatibility of methods of measurements in Soviet and modern Russian statistics; incomplete pools of data associated with various negative processes, administrative reforms and institutional changes in the Russian economy; partial loss of historical statistical data. This determines the inability to use time series models of economic indicators to identify the main development patterns, and the resulting trends are limited and not reliable enough. Furthermore, the majority of methods have applicability limits and cannot be applied to incomplete or noisy data. Therefore, it is proposed to use neural network technologies and intelligent information systems developed on their basis as alternative approach for solving semistructured and unformalized problems of analyzing the living standards, as well as constructing forecasting models. A comparative analysis of the research by Russian and foreign authors on the problems of assessment and constructing a composite index of the quality of life of the population is carried out. The author proposes an assessment model, presents the results of neural network construction, and gives economic interpretation of the obtained results.