Data mining for statistical analysis of money laundering transactions
Published date | 07 October 2019 |
Pages | 753-763 |
DOI | https://doi.org/10.1108/JMLC-03-2019-0024 |
Date | 07 October 2019 |
Author | Mark Eshwar Lokanan |
Data mining for statistical
analysis of money laundering
transactions
Mark Eshwar Lokanan
Faculty of Management, Royal Roads University, Victoria, Canada
Abstract
Purpose –The purpose of this paper is to use statistical techniques to mine and analyze suspicious
transactions. With the increase in money laundering activitiesacross various sectors in some of the world’s
leading democracies,the ability to detect such transactionsis gaining grounds with more urgency. Regulators
and practitionershave been calling for an approach that can mine the large volumeof unstructured data form
suspiciousmoney laundering transactions to inform publicpolicies.
Design/methodology/approach –By deducing from the results of empirical studies in the field of
money laundering detection, this paper presented an overview of data mining technology for detecting
suspicioustransactions.
Findings –After chronicling the data mining process, the paper delvesinto an analysis of the statistical
approaches that can be used to differentiate between legitimate and suspicious money laundering
transactions. The different stages of the data mining process are carefully explained in relation to their
application to anti-money laundering compliance. The results indicate that statistical data mining
methodologyis a very efficient and useful technique to detect suspicious transactions.
Practical implications –The paper is of relevance to regulators and the financial service sector. A
discussion of how datacan be mined to facilitate statistical analysis can be used to inform regulatorypolicies
on the detectionand prevention of money laundering activities in the financialservice sector.
Originality/value –The paper discuss approaches that illustrate how analysts can use statistical
techniquesto analyze data for suspicious money laundering transactions
Keywords Compliance, Money laundering, Machine learning, Data mining, Algorithm, Data analysts
Paper type Technical paper
Introduction
Up until the big data revolution, the approach to money laundering detection has been
to look at evidence provided from audit trails and the wider business context. Such
investigative techniques mostly focused on detecting suspicious patterns from the data
available to identify laundered activities. The emergence of technology (such as
machine learning and algorithms) and the digital economy has changed the manner in
which data are mined for further analysis. Money laundering in major industrial
centers is becoming more complex and the sheer volume of the data is far beyond what
investigators can realistically comprehend and analyzed. For example, a recent media
report coming out of Canada noted that money-laundering activities in Vancouver,
British Columbia is estimated to be $1bn a year and is disrupting the real estate and
financial sectors (Meissner, 2019). Data analysts, despite how adept they are, simply
cannot comprehend the aggregate data available from money laundering activities
using traditional approaches. This is where data mining techniques can be useful. Data
mining techniques can allow data analysts to analyze the data and make informed
decisions on whether a particular transaction or a series of particular transactions are
Data mining
for statistical
analysis
753
Journalof Money Laundering
Control
Vol.22 No. 4, 2019
pp. 753-763
© Emerald Publishing Limited
1368-5201
DOI 10.1108/JMLC-03-2019-0024
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1368-5201.htm
To continue reading
Request your trial