Neural network as a tool to solve the problem of credit scoring

Abstract

Artificial neural networks have become one of the most actively developing areas of IT over the past decade. They were used to solve various problems in various areas of human society. In the financial sphere, the capabilities of artificial intelligence, machine learning and neural networks have also expanded the list of data analysis tools. One of the topical areas of research in the financial sector is the problem of credit scoring. Solving this problem reduces the risk of possible default, which may be caused by too many overdue loan payments. Currently, a sufficiently large range of methods has been developed to solve the problem of credit scoring, including the use of neural networks considered in the framework of the presented work. In the presented work, the application of a neural network for solving multiple regression problems is considered by the example of solving the problem of credit scoring. The article contains an analysis of influencing factors, their assessment, the choice of a model of an artificial neural network. In addition, the article describes the processes of training and testing a neural network for scoring. The developed information system should solve the problem of credit scoring, which is to determine the creditworthiness of the client. Based on the information of this client, as well as on historical data about other customers, the system determines whether the person is solvent. The field of application of the developed information system is banking, in particular retail lending.