Self-Organising Map Based Framework for Investigating Accounts Suspected of Money Laundering.

Abdallah Alshantti,Adil Rasheed

doi:10.3389/frai.2021.761925

Abdallah Alshantti, Adil Rasheed

Open Access

https://doi.org/10.3389/frai.2021.761925

Copy DOI

Abstract

There has been an emerging interest by financial institutions to develop advanced systems that can help enhance their anti-money laundering (AML) programmes. In this study, we present a self-organising map (SOM) based approach to predict which bank accounts are possibly involved in money laundering cases, given their financial transaction histories. Our method takes advantage of the competitive and adaptive properties of SOM to represent the accounts in a lower-dimensional space. Subsequently, categorising the SOM and the accounts into money laundering risk levels and proposing investigative strategies enables us to measure the classification performance. Our results indicate that our framework is well capable of identifying suspicious accounts already investigated by our partner bank, using both proposed investigation strategies. We further validate our model by analysing the performance when modifying different parameters in our dataset.

Highlights

Money laundering represents a major challenge for governments and financial institutions alike, as the flow of dirty money can hinder a state’s development, damage the reputation of the financial system and motivate the generation of further crime (Kumar, 2012)
In contrast to the other studies presented in the literature, we developed a model that detects money laundering activity in an imbalanced dataset while demonstrating robustness against the inadequately labelled data
The poor labelling is a general problem in money laundering datasets which stems from the fact that a remarkable proportion of money laundering transactions are undetected by conventional rule-based alert systems

Summary

Introduction

Money laundering represents a major challenge for governments and financial institutions alike, as the flow of dirty money can hinder a state’s development, damage the reputation of the financial system and motivate the generation of further crime (Kumar, 2012). A major drawback of the rule-based systems is the generation of a significant volume of false positive alerts that are costly in terms of time and resources needed to track down flagged cases (Gao, 2009). Those false positive alarms are estimated to constitute more than 90% of the total alerts generated by the traditional rulebased systems commonly adopted by banks (Breslow et al, 2017). There has been an increasing desire to develop more advanced tools for more precise detection of money laundering transactions

Objectives

Methods

Results

Conclusion