Neural Network Modeling Of E-Government Development And The Socio-Economic Environment

Abstract

             The article deals with the choice of indicators for modeling the economic component of sustainable development in the countries of the world. The study was conducted on the basis of a collaboration with NNC “World Data Center for Geo-informatics and Sustainable Development”. Methodical backgrounds are generalized scenario modeling of country development. The choice of key metrics for scenario modeling in four directions: finance, energy, macroeconomic environment, gross domestic product structure. The methods of scenario construction are analyzed and the feasibility of using methods based on Bayesian probabilistic models is determined. Obstacles to the construction of adequate models are indicated. Advantages of using neural network technologies to achieve the goal of the study are substantiated, in particular avoiding problems of heterogeneity of the set of initial indicators, lack of data, complexity of long-term forecasting with low level of information content and availability nonlinear relationships between indicators.