Using the method of intelligent neural networks in identifying and rating the behavior of customers suspected of money laundering) Study case: Bank Mellat(

Arjmandnezhad, AbdolMahdi; Habibi, Reza; Babaee Keyapay, Masoud

doi:10.22034/jifb.2024.451360.1567

Using the method of intelligent neural networks in identifying and rating the behavior of customers suspected of money laundering) Study case: Bank Mellat(

Document Type : Original Article

Authors

AbdolMahdi Arjmandnezhad ¹

Reza Habibi ²

Masoud Babaee Keyapay ³

¹ Central Bank Of Iran

² Faculty Member, Iran Banking Institute

³ Mellat Bank

https://doi.org/10.22034/jifb.2024.451360.1567

Abstract

Money laundering as a criminal phenomenon can cause a disruption in the functioning of banks and impose a lot of costs on them. The current strategy of most Iranian banks is to identify this category of customers using general rules, which allows many false positives and makes it difficult to detect money laundering operations. The purpose of this study is to provide criteria for identifying bank customers with the highest probability of suspicion of money laundering using smart algorithms. To this aim, a two-stage model was used to analyze customer behavior using a database of customer characteristics and their financial data in the second six-month period ending in March 1401. In this study, techniques such as self-organizing mapping (SOM) and dependency rules have been used to create a comprehensive profile of customers. Research variables include demographic data, banking services, and financial transactions of customers, extracted from the database of Bank Mellat. The results show that the research model is well able to identify and analyze the customers suspected of money laundering, already investigated by the operating bank. Using both proposed models, the results do not show a significant difference.

Keywords

Artificial neural networks

self-organizing map

behavioral scoring

risk-based approach