Author:

Marina Yu. Malkina, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia

Andrey L. Sochkov, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia

Yulia I. Kapustina, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia

Abstract:

The dynamically developing economy subjected to the influence of global uncertainty contextualises the use of artificial intelligence methods that enable the construction of advanced adaptive models based on the nonlinear interaction of variables. Such models allow creating more accurate economic forecasts and scenarios for the socioeconomic development compared to traditional econometric and statistical methods. The study focuses on the neural network modelling and forecasting of gross regional product (GRP) of a constituent entity of the Russian Federation taking the Nizhny Novgorod oblast as an example. Theoretically and methodologically the research rests on the extended Cobb–Douglas production function along with the fundamental concepts of regional economics and neural network modelling. The study uses the 2000–2023 regional and macroeconomic data from the Federal State Statistics Service, the Bank of Russia, and the online portal Investing.com. The involvement of the data on regions with similar industrial structure and economic scale has allowed expanding the dataset for the model training. As a result, the paper presents two constructed models of GRP of the Nizhny Novgorod oblast: 1) a basic one, relying on a limited number of input parameters and data from benchmark regions according to the oblast’s Development strategy; 2) an extended one, based on a larger number of input parameters and data from regions from the same cluster as the Nizhny Novgorod oblast. The obtained models allowed producing three GRP forecasts for the Nizhny Novgorod oblast for 2025–2027: a realistic, an optimistic, and a pessimistic one. The results of the realistic scenario turned out to be close to the regional government’s forecast. The extended model, based on wider databases and a larger number of input parameters, provided more accurate forecasts. The results and conclusions of the study may be useful in forecasting and managing the socioeconomic development of Russia and its regions.

Keywords: gross regional product; regional economy; economic forecast; development scenarios; artificial intelligence; machine learning; the Nizhny Novgorod oblast.

Download article

For citation: Malkina M. Yu., Sochkov A. L., Kapustina Yu. I. (2025). Forecasting regional GRP using artificial intelligence methods. Journal of New Economy, vol. 26, no. 3, pp. 67–85. DOI: 10.29141/2658-5081-2025-26-3-4. EDN: NWHFVD.