Сибирский федеральный университет

Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: International Conference on Economic and Social Trends for Sustainability of Modern Society; Krasnoyarsk; Krasnoyarsk

Год издания: 2020

Идентификатор DOI: 10.15405/epsbs.2020.10.03.130

Ключевые слова: credit scoring, support vector machine

Аннотация: Scoring is one of the main tools in the banking system to determine the creditworthiness of customers. There are many methods for processing and storing customer information, but the effectiveness of these methods is related to the quality of the data provided and the process of their processing. In this regard, there is a need to study methods and find the optimal one. This article describes the possibility of using the support vector machine (SVM) to solve the credit scoring problems. The concept and essence of the SVM are considered. The support vector machine permits to build a good classifier with a minimum of initial features. In difference with other methods, the SVM is the most optimal for determining the creditworthiness of a client. We trained the sample on one set and checked on another test sample. As a result of testing the support vector machine (50 starts), a minimum error of 15.6% was obtained. The article also proposes a structural-functional model of the loan processing system using this method. The article describes the decision-making process for granting a loan in the form of a structural-functional model in the form of a diagram. The structural and functional model involves the application of the process of processing loan applications using the support vector machine as a software module in an automated banking system.

Журнал: European Proceedings of Social and Behavioural Sciences EpSBS

Выпуск журнала: 90

Номера страниц: 1132-1139

Место издания: Krasnoyarsk

Издатель: European Proceedings

• Chuvashova M. N. (Irkutsk State University)
• Agafonov E. D. (Reshetnev Siberian State University of Science and Technology)
• Zhuravleva I. A. (Irkutsk State University)
• Polyushkevich O. A. (Irkutsk State University)

• РИНЦ (eLIBRARY.RU)