SINOPSIS

The observed trend of active development of consumer, educational, mortgage and other crediting, along with high competition in the modern crediting market, makes it necessary to pay even more attention to mathematical modeling of the assessment of the current and future solvency of natural persons, and hence the credit risk management process of a commercial bank directed to reduce losses associated with a significant increase in overdue loans. The article presents approaches to the assessment of the solvency of natural persons, based on the analysis of factors related to the relations of lenders and borrowers on the parameters of their solvency and reliability based on multi-factor expert, neural network and fuzzy types of modeling. The developed tools can serve as the basis of a management decision support system in financial institutions, it is distinguished by the ability to highlight reliably the characteristics of a potential client related to the high-risk area. A method of balanced multicriteria credit assessment of natural persons is proposed within the framework of these tools, which provides for the compilation of expert estimates regarding the priority of indicators of creditworthiness, in general, and current indicators of the solvency of natural persons, in particular. The compilation of acquired expertise is carried out using neural network modeling and the fuzzy maximin convolution method of qualitative criteria for assessing the solvency of natural persons. Selected assessment criteria of natural person solvency are differentiated by degrees of their priorities, which at the initial stage of the examination are identified based on an agreed opinion of specially invited experts in the form of generalized weights of their relative influence on the solvency level of the natural persons.