search for




 

Study on the real owner distinction in financial transactions using deep learning
Journal of the Korean Data & Information Science Society 2021;32:781-97
Published online July 31, 2021;  https://doi.org/10.7465/jkdi.2021.32.4.781
© 2021 Korean Data and Information Science Society.

Seung Eon Lee1 · Hyun Se Kim2 · Nam Ho Kim3 · Jung Yun Um4 · Ji Hwan Woo5

12345AI Competency Center, Shinhan Bank
5Korea University Graduate School of Management of Technology
Correspondence to: 1 AI Competency Center, Shinhan Bank, 55, Sejong-daero, Jung-gu Seoul, Koera, 04513, Korea.
2 AI Competency Center, Shinhan Bank, 55, Sejong-daero, Jung-gu Seoul, Koera, 04513, Korea.
3 AI Competency Center, Shinhan Bank, 55, Sejong-daero, Jung-gu Seoul, Koera, 04513, Korea.
4 AI Competency Center, Shinhan Bank, 55, Sejong-daero, Jung-gu Seoul, Koera, 04513, Korea.
5 AI Competency Center, Shinhan Bank, 55, Sejong-daero, Jung-gu, Seoul 04513, Koera, Adjunct professor, Korea University Graduate School of Management of Technology.
Email: jihwan woo@korea.ac.kr
Received June 26, 2021; Revised July 13, 2021; Accepted July 15, 2021.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
Determining whether the actual party to a transaction is a legal entity or a natural person in financial transactions is a key aspect of the customer identification system. However, the existing deep learning-based natural language processing technology has a disadvantage in that it cannot be applied to determine whether a given word is a name of a natural person or a corporate because it recognizes a word or sentence as a minimum unit of natural language. In this paper, based on deep learning, a model for judging whether a given word is a natural person’s name was proposed for the first time in the financial sector through analysis of syllables, the smallest unit constituting a word in Korean. Through the proposed method, it is possible to distinguish between legal entities and natural persons in financial transaction data with a high level of accuracy and speed. Therefore, it is possible to replace the work that relied on the experience of professional personnel when performing the existing financial services with artificial intelligence, and the throughput is expected to increase. In addition, it is meaningful in that it provided a benchmark that can be referenced when introducing this method in the financial industry in the future by providing a comparative experiment on the customer confirmation model with various AI-based models.
Keywords : Customer due diligence, digital finance, language Model.