Open AccessUSING MACHINE LEARNING METHODS TO ASSESS RISKS WHEN IMPLEMENTING A NEW CREDIT PRODUCT
Author(s) -
S. P. Bobkov,
Станислав Вадимович Суворов,
Артем Игоревич Орлов,
Егор Алексеевич Пивнев
Publication year - 2020
Publication title -
izvestiâ vysših učebnyh zavedenij. seriâ "èkonomika, finansy i upravlenie proizvodstvom"/izvestiâ vysših učebnyh zavedenij. seriâ «èkonomika, finansy i upravlenie proizvodstvom»
Language(s) - English
Resource type - Journals
eISSN - 2713-1114
pISSN - 2218-1784
DOI - 10.6060/ivecofin.2020464.509
Subject(s) - novelty , loan , computer science , product (mathematics) , credit risk , credit rating , risk analysis (engineering) , machine learning , business , finance , philosophy , geometry , theology , mathematics
The article discusses the issues of assessing the creditworthiness of individuals using credit scoring. This rating system is an effective approach to determining the level of risk for a specific customer segment. This is especially true of the situation when a credit institution launches a new credit product. The main idea proposed in the article is that new customer scoring cards are created on the basis of existing cards by mathematical data processing. The novelty of the method lies in the fact that the scoring is done based on a dedicated subset of customer data stored in the corporate storage. The approach helps to make a decision on granting a loan and can be recommended for use in lending