|

The use of neural network technologies in the field of consumer lending

Authors: Askerova N.A., Sokolov T.D., Askerova A.A.
Published in issue: #5(58)/2021
DOI: 10.18698/2541-8009-2021-5-695


Category: Informatics, Computer Engineering and Control | Chapter: System Analysis, Control, and Information Processing, Statistics

Keywords: neural networks, technology, financial sphere, methods, creditworthiness, libraries, efficiency
Published: 20.05.2021

The paper considers the application of neural networks in assessing the creditworthiness of a borrower, gives examples of their implementation in this area, describes the advantages and disadvantages of neural networks in comparison with expert assessment in the lending issue. A comparison of the creditworthiness of an individual is carried out using an expert approach and using various libraries of neural networks according to the following criteria: verification time, training time, financial costs, characteristics of the processed data and accuracy. Authors formed a table of indicators for comparison and assessed the effectiveness of using neural networks to determine the creditworthiness of individuals.


References

[1] Kak rabotaet neyronnaya set’: algoritmy, obuchenie, funktsii aktivatsii i poteri [How neural network works: algorithms, learning, activation functions and losses]. neurohive.io: website (in Russ.). URL: https://neurohive.io/ru/osnovy-data-science/osnovy-nejronnyh-setej-algoritmy-obuchenie-funkcii-aktivacii-i-poteri/ (accessed: 2020.05.07).

[2] Ekspertnaya otsenka. biznes-prost.ru: website (in Russ.). URL: https://biznes-prost.ru/ekspertnaya-ocenka.html (accessed: 2020.05.07).

[3] Mnogosloynaya neyronnaya set’ [Multilayer neural net]. wiki.loginom.ru: website (in Russ.). URL: https://wiki.loginom.ru/articles/multilayer-neural-net.html (accessed: 2020.05.07).

[4] Gusarova O.M., Komarov P.I., Denisov D.E. Neural networks in credit scoring. Vestnik Altayskoy akademii ekonomiki i prava, 2018, no. 5, pp. 107–113 (in Russ.).

[5] Wallace M.P. Neural networks and their application to finance. In: Operations management. Apple Academic Press, 2016, pp. 67–76.

[6] Prosalova, V.S. Problemy otsenki kreditosposobnosti klientov kommercheskikh bankov [Problems of assessing client’s creditworthiness in commercial banks]. Vladivostok, Dal’nauka Publ., 2008 (in Russ.).

[7] Klassy kreditosposobnosti Sberbanka [Classes of Sberbank creditworthiness]. ekamoney.ru: website (in Russ.). URL: http://ekamoney.ru/klassyi-kreditosposobnosti-sberbanka (accessed: 2020.05.07).

[8] Kruchinin D., Dolotov E., Kustikova V., et al. Sravnenie bibliotek glubokogo obucheniya na primere zadachi klassifikatsii rukopisnykh tsifr [Comparing deep learning libraries at the example of numeric handprinting classification]. habr.com: website (in Russ.). URL: https://habr.com/ru/company/intel/blog/254747 (accessed: 2020.05.07).

[9] Deep learning na pal’tsakh: mozhet li obychnyy chelovek izuchit’ neyronnye seti i khorosho na etom zarabotat’ [Deep learning in layman’s terms: if common person can learn neural network and make good money with it]. ostrovrusa.ru: website (in Russ.). URL: https://ostrovrusa.ru/deep-learning-na-palcah (accessed: 2020.05.07).

[10] Skol’ko poluchayut i zarabatyvayut sotrudniki Sberbanka? [How much do Sberbank workers get and earn?] skolkozarabativaet.ru: website (in Russ.). URL: https://skolkozarabativaet.ru/biznes/finansovye-novosti/skolko-poluchayut-i-zarabatyvayut-sotrudniki-sberbanka (accessed: 2020.05.07).