We are in social networks:
Vkontakte Facebook Twitter

Assessment of possible bankruptcy based on the indicators of the financial state of the companies using counterpropagation neural networks

A. D. Sharapov, D. B. Kaydanovych

There is developed in the article the new model of analysis of the financial state of enterprises and assessment of its bankruptcy risk. The proposed model is based on the tools of artificial counterpropagation neural network which consists of Kohonen and Grossberg layers of neurons. For estimation of possibility of bankruptcy the companies are divided into two groups — bankrupts and financially stable enterprises. It gives the opportunity to expose the key characteristics and specified financial indexes for these classes. Carried out experiments demonstrated the high efficiency of the developed models.

Keywords. Bankruptcy diagnosis, artificial neural network, counterpropagation neural network, Kohonen map, Grossberg layer of neurons, self-organization.

Download publication