Prytomanova O. Commercial Bank Credit Risk Modeling on the Basis of the Neural-Fuzzy Technologies

The object of research is the crediting process in the activity of commercial bank with the view on risks. The aim is to create the system of decision-making support for the analysis and the effective management of the commercial bank borrower's credit risk by implementation of adequate mathematical models built on the basis of the neural-fuzzy technologies. Methods - statistical, econometrical, theories of fuzzy multitude and neural-fuzzy nets, system analysis. Results - new neural-fuzzy models have been elaborated, i.e. the assessments of borrower's solvency indices influence on the level of connected commercial bank credit risk as well as forecasting of borrower's financial activity indices dynamics which allow more adequately reconstructing complex non-linear dependences of the credit risk level from the main borrower's solvency factors under the condition of fuzzy both qualitative and quantitative statistical information under the relatively small experimental samplings. The elaborated models wereimplemented in the activity of credit and financial administration "Credit-Dnepr" under the construction of the system of risk evaluation. Field - financial activity.