About the issue
Publications
Partners
September 2022 (published: 27.09.2022)
Number 3(50)
Home > Issue > Expert method and neural networks for assessing the risk of bankruptcy
of an enterprise
Shimokhin A.V. , Boltovsky S.N.
In the modern world, IT technologies have firmly taken root in the business processes of an organization, which allow the staff to be freed from routine operations, and the latest achievements of IT technology are even more intelligent. Basically, the development of intellectual support for business processes occurs due to the develop-ment of such technologies as, for example, more data, neural networks and others. Of particular interest are neural networks that have the ability to self-learn using the experience of experts in solving a particular problem in business, marketing, etc. The advantages of neural networks over classical linear and nonlinear methods of statistics are described. The range of management tasks is considered, in which neural networks can provide in-telligent support in matters of supplier selection, outsourcing, bankruptcy risk assessment. A technique for train-ing a neural network that performs intellectual support in solving management problems is proposed. Known models for assessing the risk of bankruptcy of various types of enterprises are given. Based on these models and the proposed methodology, a neural network was developed and trained, which, taking as input the values of working capital, assets, net profit, equity and revenue for a certain period, can predict the risk of bankruptcy. And classify the business as financially sound, unstable, or at high risk of bankruptcy. As a result of the creation and training of a neural network for intelligent support in predicting the risk of bankruptcy of an enterprise. The resulting neural network consists of 2 hidden layers, the first contains fifty neurons, the second ten. A graph of this network is provided. The hyperbolic tangent was used as the activation function
Read the full article
Keywords: neural network modeling, bankruptcy risk, Python, bankruptcy risk assessment models.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
UDC 330.46
Expert method and neural networks for assessing the risk of bankruptcy
of an enterprise
In the modern world, IT technologies have firmly taken root in the business processes of an organization, which allow the staff to be freed from routine operations, and the latest achievements of IT technology are even more intelligent. Basically, the development of intellectual support for business processes occurs due to the develop-ment of such technologies as, for example, more data, neural networks and others. Of particular interest are neural networks that have the ability to self-learn using the experience of experts in solving a particular problem in business, marketing, etc. The advantages of neural networks over classical linear and nonlinear methods of statistics are described. The range of management tasks is considered, in which neural networks can provide in-telligent support in matters of supplier selection, outsourcing, bankruptcy risk assessment. A technique for train-ing a neural network that performs intellectual support in solving management problems is proposed. Known models for assessing the risk of bankruptcy of various types of enterprises are given. Based on these models and the proposed methodology, a neural network was developed and trained, which, taking as input the values of working capital, assets, net profit, equity and revenue for a certain period, can predict the risk of bankruptcy. And classify the business as financially sound, unstable, or at high risk of bankruptcy. As a result of the creation and training of a neural network for intelligent support in predicting the risk of bankruptcy of an enterprise. The resulting neural network consists of 2 hidden layers, the first contains fifty neurons, the second ten. A graph of this network is provided. The hyperbolic tangent was used as the activation function
Read the full article
Keywords: neural network modeling, bankruptcy risk, Python, bankruptcy risk assessment models.
DOI 10.17586/2310-1172-2022-17-3-156-163