Author(s): Olga V. Kitova, Ludmila P. Dyakonova,Vladimir A. Kitov and Victoria M. Savinova
Aim of the study: In the world and in Russia, the development of a digital state and a digital economy is relevant. Within the digital state, it is important to create a service for predicting socio-economic indicators based on modern methods and models of econometrics and machine learning, and make it accessible and convenient to use. The aim of this research is to develop requirements for a digital scenario forecasting service, to study possibilities of its realization on the basis of information-analytical system "SHM Horizon", developed by the authors at the Plekhanov Russian University of Economics. Methodology: This study analyzes the needs for scenario forecasting of socio-economic indicators at the federal, regional, municipal and corporate levels. The main functional and nonfunctional requirements for such a service and its software architecture are developed. The analysis of the information-analytical system "SHM Horizon" is carried out, approaches to its development are proposed to improve its models and methods using a system of multiple linear regression equations, a multilayer perceptron, regression decision trees and a neuro-fuzzy inference system ANFIS and the implementation on this basis of the scenario service forecasting social and economic indicators of the Russian Federation. Conclusion: The theoretical significance of the obtained results lies in the development of the theory of constructing hybrid systems for scenario forecasting of interrelated time series with structural shifts, which include time series of socio-economic indicators. The practical significance lies in the development of requirements for the scenario forecasting service and its architecture.