Subscribe

Going beyond simple rules with advanced analytics

White collar crime is on the increase, with cyber-crimes in particular not only growing in number, but also in breadth.


Johannesburg, 29 May 2017
William Lawrence, Regional Practice Lead: Fraud, SAS South Africa.
William Lawrence, Regional Practice Lead: Fraud, SAS South Africa.

White collar crime is on the increase, with cyber-crimes in particular not only growing in number, but also in breadth.

The targets for white collar crimes are becoming more diverse, with the scope now extending far beyond the typical credit card fraud or financial institutions' secure data.

Stopping these hackers will take more than just the frontline security solutions like firewalls and anti-malware tools. A second line of analytics-based solutions offers new ways to catch these criminals. Up until recently, this second line of defence was built using rules-based analytics tools.

"Rules-based solutions are built around 'if-then' statements, based on common sense and conventional wisdom. So - for example - if a bank were to receive an unexpected request to authorise the spending of a large amount of money in Malaysia from a South African client's credit card, it will quite likely be denied. This is because the rules would decide that because the client is South African and the transaction is occurring halfway around the world, there is a strong possibility it is criminal in nature," says William Lawrence, Regional Practice Lead: Fraud at SAS.

"From a basic security standpoint this offers some benefits, but because rules-based analytics can only provide an 'either-or' option, they do not allow systems to probe deeper into the situation, or to utilise additional data to intelligently decide on individual instances on a case-by-case basis.

He adds that another major challenge posed by rules-based analytics is that as time passes, an increasing number of rules are required to keep the system effective. As more rules are added, there is a higher probability that those rules will conflict, leading to the creation of exceptions. This, in turn, makes the entire analytics process increasingly complicated.

"This is why SAS has focused instead on advanced analytics, which are tools that are able to learn and adjust their decisions based on new information. This means that as new data is consumed, or changes to existing data are understood, the solution will adjust to and factor in these updates."

"Advanced analytics work because they are able to factor additional information into any decision that is taken. Continuing the credit card example from above, an advanced analytical solution would be able to factor in the knowledge that two days earlier, the client had purchased an airline ticket to Kuala Lumpur using their cheque account, indicating that the transaction being requested is actually genuine."

Moreover, when it comes to security, the more source data that is available about an individual, the better the results will be, states Lawrence. Continuing the banking security analogy, advanced analytics tools would, in principle, be able to collate data obtained from a vehicle tracker, from a customer's cell phone or from any number of other Internet-enabled sensors. It could then use this information to identify exactly where they were, in order to determine whether it was actually them making the purchase request in Malaysia.

Obviously, there remain privacy issues around the collection of such data, not to mention the siloed approach many businesses, including banks, adopt towards different business units. Nonetheless, advanced analytics clearly offers so much more than rules-based tools ever could.

"There can be little doubt that it has the power to open up a new security frontier, provided the organisations using these tools can strike the balance between customer desires - such as improved security - and their fears, such as enterprises having access to too much of their personal information."

"Ultimately, of course, the reality is that when it comes to tackling cybercrime, there is no silver bullet. However, when one considers that rules-based analytics can only make decisions based on past experiences, while advanced analytics are capable of both understanding the present and providing insight about what will happen in the future, it is clear that only one of these has the capability of making inroads into the ongoing battle against a growing range of cyber-threats," he concludes.

Share