Neural Network Approximation in Forecasting Economic Risks

Abstract

The article aims to assess the level of the economic systems development in the context of the accuracy of forecasts in unpredictable socio-economic conditions, taking into account the impact of unforeseen environmental risks and disasters. The authors used methods of neural networks in order to evaluate the sustainable development process. To date, a large number of mathematical forecasting methods are known, and experts in the world economy use appropriate risk assessment criteria, but the neural network is used when the exact type of connections between inputs and outputs is unknown, which allows us to create a more accurate and flexible forecast model. The modeling takes into account the main weights that determine the degree and  the priority of the impact on each component of the economic system and characterizes the complex macroeconomic relationships to determine the aggregate indices. The developed model is used when assessing the level of economic security of the national economy. Especially, such an important component as environmental security is evaluated with the help of this model