A Time-Frequency Based Suspicious Activity Detection for Anti-Money Laundering

Abstract

Money laundering is the crucial mechanism utilized by criminals to inject proceeds of crime into the financial system. The primary responsibility of the detection of suspicious activity related to money laundering is with the financial institutions. Most of the current systems in these institutions are rule-based and ineffective (over 90 % false positives). The available data science-based anti-money laundering (AML) models to replace the existing rule-based systems work on customer relationship management (CRM) features and time characteristics of transaction behaviour. Due to thousands of possible account features, customer features, and their combinations, it is challenging to perform feature engineering to achieve reasonable accuracy. Aiming to improve the detection performance of suspicious transaction monitoring systems for AML systems, in this article, we introduce a novel feature set based on time-frequency analysis, that uses 2-D representations of financial transactions. Random forest is utilized as a machine learning method, and simulated annealing is adopted for hyperparameter tuning. The designed algorithm is tested on real banking data, proving the results' efficacy in practically relevant environments. It is shown that the timefrequency characteristics are discriminatory features for suspicious and non-suspicious entities. Therefore, these features substantially improve the area under curve results (over 1%) of the existing data science-based transaction monitoring systems. Using time-frequency features alone, a false positive rate of 14.9% has been achieved, with an F-score of 59.05%. When combined with transaction and CRM features, the false positive rate is 11.85%, and the F-Score is improved to 74.06%. INDEX TERMSAnomaly detection, anti-money laundering, compliance, random forest algorithm, timefrequency analysis, transaction monitoring.

Predicting Suspicious Money Laundering Transactions using Machine Learning Algorithms

“…However, this is only true when SARs are used with additional measures such as transaction monitoring, CDD, and KYC (Singh and Best 2019;Lokanan 2022;Zavoli and King 2021). This underscores the need for multiple layers of defense to stay ahead of money launderers' attempts to infiltrate financial systems (Lokanan 2022;dalla Pellegrina et al 2020;Ketenci et al 2021). Studies have also shown that the accuracy and reliability of SARs are highly dependent on the knowledge and expertise of those preparing them (Zavoli and King 2021;Loh 2021).…”

“…With globalized banking systems, the movement of capital across different countries, and sophisticated money laundering techniques, compliance with AML regulations has become an increasingly pressing topic (Mao and Dawod 2022;Masciandaro and Filotto 2001). The increasing challenges of detecting suspicious transactions has led to a greater reliance on SARs (Ketenci et al 2021;Singh and Best 2019). Previous studies tracking the effectiveness of SARs suggest that they are a useful tool for law enforcement and financial institutions to mitigate the threat of money laundering (Ketenci et al 2021;Singh and Best 2019;Coombs-Goodfellow and Lokanan 2018).…”

“…The increasing challenges of detecting suspicious transactions has led to a greater reliance on SARs (Ketenci et al 2021;Singh and Best 2019). Previous studies tracking the effectiveness of SARs suggest that they are a useful tool for law enforcement and financial institutions to mitigate the threat of money laundering (Ketenci et al 2021;Singh and Best 2019;Coombs-Goodfellow and Lokanan 2018).…”

“…For example, regarding CDD, ML has been used to analyze customer data to identify high-risk customers or transactions that may indicate money laundering or other financial crimes (Masciandaro and Filotto 2001;Kurum 2020). In other areas such as transaction monitoring, AI has analyzed real-time transactions to identify unusual or suspicious money laundering activity (Han et al 2020;Kurum 2020;Ketenci et al 2021). AI solutions in AML compliance have increased efficiency and accuracy in predicting money laundering activities (Lokanan 2022;Chen et al 2018;Turki et al 2021;Canhoto 2021).…”

Money Laundering Detection with Node2Vec

“…Money laundering activities may be detected with different kinds of solutions apart from graph-based machine learning methods. For example, the average, variance, kurtois, sparsity, discontinuity values, which are calculated with very known formulas, and transaction information obtained by examining bank accounts in the time and frequency domains may be used as features in machine learning algorithms to detect money laundering [19].…”