High-performance computing, combined with analytical techniques, behavioural analytics and investigative workflow, can help combat fraud in the digital era, says SAS Institute, a software analytics leader.
"The rapid advance of technology is empowering banks and financial institutions with fraud analytics to help them detect threats more efficiently in real time," said David Stewart, business director of SAS Institute.
David Stewart, business director of SAS Institute
Mr Stewart said high-performance computing makes analytics more practical, allowing the analysis of massive amounts of relevant data in full context.
For example, using in-memory analytics, simulations that once took hours take just seconds.
Fraud analysts can rapidly test multiple methods to determine which models work best.
Affordable high-performance computing helps make new investments in high-performance analytics profitable, Mr Stewart said.
He said a combination of analytical techniques can allow more insight into data by detecting and guarding against fraud with fewer false positives.
Neural network models use machine learning to interactively learn from the data without human intervention, allowing algorithms to evolve and become more accurate with every iteration.
Behavioural analytics transcend rules-based systems since they help capture behavioural patterns from every source, and evaluate information every time a transaction occurs.
This process builds detailed profiles of accounts, cardholders, merchants, point-of-sale terminals, devices and web sessions.
The higher the number of profiles available, the richer the understanding of whether a payment transaction or product application is legitimate, said Mr Stewart.
"The goal of behavioural analytics will be to identify potentially fraudulent behaviour before a payment is made or a customer's account is compromised," said Mr Stewart.
Fraudsters can easily outmanoeuvre existing rules-based systems, so it's essential to have adaptive analytics that can detect unknown risks, he said.
Mr Stewart said the investigative workflow becomes more efficient in detecting, triaging and building investigations into suspicious activity.
Firms need transaction data that can be shared to associated accounts, and from those accounts to other parties: households, corporate parents or other networked entities.
Aggregate work items automatically assemble alerts from multiple monitoring systems, by prioritising higher-risk activities, rather than require analysts to review work items independently.
"While the government has implemented policy to digitise the payment process, more education is needed for people to understand its value and its threats," said Mr Stewart.
He said financial crime is becoming more global and sophisticated. Investment in IT security is of paramount importance to secure data and prevent attacks.
Banks and financial institutions will need to invest in financial intelligence to understand risks in more detail and build a clearer picture of how criminals operate.
Analytics plays a role more for fraud management and prevention as it enables banks to develop a comprehensive view of how criminals seek to abuse the financial system.
Instead of treating a suspicious customer or transaction in isolation, new tools can identify the connections between suspicious accounts within a network and see how money flows between them and other networks.
This provides valuable insight that informs current and future risk management decisions.
He suggested bank and financial institutions to centralise their fraud management operations and constantly update fraud management systems with new rules, statistical models and acquired knowledge.
Mr Stewart said the shortage of skilled labour in analytics is increasing globally.
SAS predicts that demands for data science professionals is expected to double over the next three years, particularly for startups and SMEs.