Anti-fraud technologies get smarter

With card fraud and other forms of fraud reportedly on the rise during the economic downturn, anti-fraud management software vendors are having to up their game to play catch-up to the fraudsters.

Business intelligence and analytics vendors such as SAS, focus on not just looking at fraud in terms of monitoring card transactions, but the ability to match seemingly unrelated events across different parts of the business. Its real-time card fraud detection system which is used by banks such as HSBC, reviews card transactions alongside other changes in customer behaviour and then based on that analysis advises in real time as to whether a card transaction should proceed or be flagged for further investigation. Using such a system, HSBC claims to have reduced false positive rates, which is one of the biggest bug bears of any fraud detection system.

But while banks may deploy a system to try and combat the different forms of fraud that are prevalent today, it needs to be flexible enough to predict and detect changes in fraudsters' behaviour patterns in order to avoid detection. Analytics and decision management vendor, Fair Isaac Corporation, is trying to do this in the debit and credit card space with version 6.0 of its Falcon Fraud Manager scoring server, which uses recent advances in fraud analytics and profiling to help banks more quickly identify changing fraud patterns.

Using what it calls, "adaptive analytics", Fair Isaac provides "dynamic, real-time self-calibration of fraud detection models" to help firms more quickly identify changing fraud behavior patterns and improve fraud detection performance.

Another software provider, Actimize has incorporated IBM InfoSphere's Global Name Recognition (GNR) technology into its risk management platform. GNR is designed to help firms overcome challenges in matching names across different cultural and language barriers as part of their financial crime fighting efforts and works by analysing the order of a name, cultural spelling variations, nicknames and different spelling variations.

"[Global Name Analytics] ... can help to identify and correct names that may have been presented in non-standard or incorrect sequential order." It can also identify the "cultural classification" of a person's name using linguistic and statistical tools, which can be useful for Know Your Customer regulations and anti-fraud projects that entail matching names against lists or databases of suspected terrorists or Politically Exposed Persons.