Forecasting indicators of territorial entity based on phenomenological models of collective behavior

Abstract

The article is aimed at solution of the urgent scientific task of creating new methods of forecasting analytics that allow obtaining more accurate and reliable forecasts of the socioeconomic development of countries, regions and cities. In contrast to traditional methods it is achieved by simultaneously taking into account both the dynamic patterns of changes in the states of individual objects, and the statistical patterns of their collective behavior. It is shown that such forecasting methods can be based on the use of phenomenological approaches and techniques in relation to the description of statistical data and states of objects in multidimensional spaces of social and economic indicators. This scientific direction refers to the current section of system dynamics and sociophysics and allows to formulate a fundamentally new approach to modeling and forecasting the development of territorial entities, based on the application of the concept of continual fields of empirical measures and various models of object state spaces. On the basis of considering statistical distributions of joint events observation of a set of indicators there have been proposed dependencies for predicting the states of objects. As an implementation of the developed method, forecasting of some indicators of the development of cities of Russia was carried out.