Creation of an economic risk management system using artificial neural networks

• Main Authors: Sleptsova Yulia, Kachalov Roman, Shokin Yan
• Format: Article
• Language: English
• Published: Peter the Great St. Petersburg Polytechnic University 2020-10-01
• Series: Научно-технические ведомости СПбГПУ: Экономические науки
• Subjects: risk level managerial system; intentional; expectational; cognitive and functional subsystems; risk factors; anti-risk impact actions; artificial neural networks
• Online Access: https://economy.spbstu.ru/article/2020.85.02/

Due to the abundance and fragmentation of information in the digital economy, the risk management system of the enterprise and its socio-economic ecosystem should rely on such digital technologies, which can be used to gain time to assess and analyze changes in the economic environment. The purpose of this article is to formulate basic approaches to creating a risk level managerial system, including processes for identifying, evaluating and minimizing the risk level in the preparation of management decisions developed using artificial neural networks. This study shows using the methods of operational risk management theory, system economic theory, algorithm theory, in particular, artificial neural networks, and immune response modeling, that the risk management system of a modern enterprise and its socio-economic ecosystem will be based on the principles of functioning of the immune system by analogy with similar systems in living organisms. Economic risk management processes are modeling within four main interacting subsystems: intentional, expectational, cognitive, and functional. The principles that must be observed for the correct use of artificial neural networks in the preparation of management decisions and for the accumulation of information about the level of possible risk are highlighted. For wide application of artificial neural networks in enterprises, it is necessary to reach a certain level in digital technologies. It is shown that to create a specialized operating system for managing the risk level of an industrial Internet of things (IoT) enterprise or a digital multi-party business platform as a whole, a separate digital ecosystem may be required. The presented research may be useful for specialists and managers of enterprises when creating risk management systems and management decision support systems using artificial neural network algorithms. The lack of development of the "basic" level of information technologies at the enterprise is a limitation of the application of the results obtained.