Analysis of Modern Credit Risk Assessment Models for the Russian Market

An analysis of modern credit risk assessment models (expert systems, scoring, market-based and neural network models) was conducted, identifying their advantages and limitations. It is shown that in the Russian context, market-based models are less applicable due to the underdeveloped financial market, while neural networks and fuzzy logic hold promise for accounting for nonlinear dependencies. The need to adapt foreign approaches and integrate new technologies, including AI, to improve assessment accuracy is substantiated. The study is based on a comparative analysis of scientific publications and banking practices.