home icon

Applicability of Big Data for Effective Anti-Money Laundering

Comments ,(0)

The scope of 'Financial-crime' has considerably widened over time, due to which Governments and Organizations alike need to be extremely cautious; anti money laundering, anti-fraud, anti-corruption, sanctions and embargoes are they key. Anti-Financial crime and Money Laundering is a big area of spending, especially for financial institutions. As per a leading analyst firm, Risk & Compliance spending will grow from USD 79 to USD 97 Billion globally, with a significant amount being spent specifically on compliance domains. From a business perspective, the major areas of Regulatory & Compliance focus include SEC, FINRA, FINCEN, FCA, OCC, APRA, FINMA, RBI / SEBI and MSA to name a few.

How do Financial Institutions successfully identify the legitimacy of the millions (and in some cases billion!) of financial transactions occurring every day? The answer- A robust Anti Money Laundering Strategy powered by Big Data.

How does it work?

An Anti-Money Laundering Big Data engine collects the raw, external data from various sources such as Know Your Customer (KYC) information, real time transaction data, regulatory data etc. The input data undergoes enrichment, transformation, and vectorization, post which it is evaluated and scored for fraud checks. Event data often needs to be combined with data from other sources such as location, account details, or transaction data from other systems prior to being evaluated for norms such as security intelligence. AML engine uses high volume data inputs, click stream data, combination of rule based models, dynamic profiling analytics, intelligent scoring algorithms and Dynamic Anomaly Detection rules for fraud scoring & investigation. From AML engine, Risk management systems & Regulatory reporting identify the Transaction risks and compliance risks.

What is an ideal Big Data Platform, best suited for Anti Money Laundering?

The Hadoop Big Data platform possesses the critical attributes ideal for AML activities. Some of these include pulling data; structured as well as unstructured, from various sources with ease, preparing the data followed by cleansing (this is critical for accuracy of insights), building a data model for analysing the data and compliance checks. The AML architecture is fully integrated with an organization's data hub. Staging Data will be complex since it contains multiple source data and this staging data provides runtimes for the predictive models to perform fraud detection.

Machine Learning (Neural Networks, Decision Trees, Bayesian Analysis etc.) is another key lever to predict and prevent Money Laundering patterns in Financial Institutions, by analysing an identified set of illicit operations. The system can also be taught to differentiate fraudulent transactions from legitimate ones by analysing data base(s) having records of only legitimate transactions.

About Author

Venkatesan Marimuthu- Principal Consultant, Testing Services, Wipro, Ltd.

Venkatesan Marimuthu is working at Wipro since June 2006 as a Principal Consultant, having 19+ years of working experience in DWH development, Project Management, Banking Domain, ETL Testing and Solution consultation

Read all blogs

Comments (0)

Post Comments