Mathematical Modeling of Credit Scoring System Based on the Monge-Kantorovich Problem

Abstract

In this article the issues of further improvement of the credit system in order to ensure the implementation of the Decree of the President of the Republic of Uzbekistan dated May 12, 2020 No 5992 "On the Strategy of Banking Reform in the Republic of Uzbekistan for 2020-2025" studied. Problems of information security in the credit system and their solutions, issues of secure storage of personal data of borrowers through blockchain technology, as well as effective modeling of lending and borrower relations, lending institutions on the basis of established criteria recalculation model and quantitative analysis of the assessment, the application of the Monge-Kantorovich problem in the development of the credit scoring model, the process of forming a credit scoring model based on the model applied to the Monge-Kantorovich transport problem. The credit scoring model based on the Monge-Kantorovich problem is similar to the Bayesian network-based credit scoring model, and calculations are performed using elements of probability theory. Borrowers were evaluated using optimal evaluation criteria. As a result, the process of deciding whether or not to grant a loan to a borrower has been completed. The results obtained were analyzed and a credit scoring model based on the Monge-Kantorovich problem was proposed.