Big data fights fraud
As the hype around big data dies down, real-use cases are beginning to surface.
For Flavio Villanustre, VP of information security at LexisNexis Risk Solutions, one of these real-world cases is the use of big data to combat a rise in fraud cases, which have emerged as part of the digital age. A company that provides data analytics services to corporate customers, LexisNexis is involved in the aggregation of information to create meaningful data, having worked with various insurance companies across the US.
According to Villanustre, integrating this data in a central data warehouse makes it easier to use, easily accessible and proves valuable for most companies. In the past, people used to think all data was complete and clean, with a unique identifier. "This is a lie. Most data is incomplete; it might not be useful and often doesn't have a key identifier. This makes the data management problem more complex than we originally thought," he said, adding that the big data problem needs to be tackled in a different way.
According to Villanustre, applying data transformation analytics to the information makes the process more efficient. Data linkage is also important, said Villanustre, as it allows the company to determine how the data is linked using probability.
LexisNexis used data-linking technologies to help a US insurance company identify fraudulent car accident claims. The company had noticed a number of shady accident insurance claims, said Villanustre, but as they were made by seemingly unrelated individuals, the company did not have any evidence to query them. "Monitoring the transactions of a single individual may not be enough. But if you look into the social group the individual is in and correlate activity across that social group, you are able to identify collusion."
LexisNexis used data-linkage technology to make sense of this data and determined the proximity of the entities making the suspicious claims. "We discovered that each one of the claims was made by people across two families who were colluding to defraud the insurance company," he said, noting that, once this was proven, the company was able to reclaim some of the money that had already been paid out.
Crime used to be much riskier, said Villanustre. "Criminals are getting smarter. They are assessing the risks and realising that there is a lot less risk involved in doing cyber crime than there is in going out and physically stealing something. And a criminal with enough knowledge and enough time can potentially create a lot more damage than if they went out onto the streets and tried to steal everything they could find."
When it comes to using big data to combat these threats, Villanustre noted that many companies lack the adequate skills and know-how to really take advantage of big data. "Yes, they have a significant volume of data that they could tap into to fight fraud, but for many companies, big data is an afterthought," said Villanustre, stressing that this is the wrong approach.
"Big data is here to stay," he concluded. "It is a powerful enabler and provides the potential that gives companies the ability to know their customers better, to understand and predict trends, and to make meaningful correlations."