The ever growing body of digital data has become an ideal camouflage for activities of both criminals and consumers.
Business intelligence (BI) tools employing social network analysis and predictive modeling features, however, can help various organizations uncover potential profits as well as hidden criminal alliances.
In the banking and insurance industries, for instance, money launderers and claims scammers have long relied on “dirty data” and human nature to mask their operations. However, a more powerful breed of data mining and analytical tools are peeling away the layers of data and uncovering “unobvious networks” to eventually reveal the identity of criminal organizations.
Earlier this year the United Kingdom’s Insurance Fraud Bureau broke an organized crime insurance fraud ring and arrested 74 people with the help of a data analytics tool developed based on SAS Institutes business intelligence software and powered by a network modeling and data analytic application from Detica Ltd., a U.K.-based software and consultancy firm.
When laundering money, organized crime rings typically hide funds under multiple accounts and transactions that appear to have been initiated by unconnected individuals or companies. Syndicates perpetrating fraudulent insurance claims employ the same tactics.
The challenge for the investigator is to discover the hidden connections that separate these transactions from activities generated by legitimate individuals and businesses.
For example, a money launderer might have opened a checking account under one name and another account using a variation of his last name. Investigators must establish links between the two accounts.
In insurance fraud, similar claims emanating from a person’s different aliases but with links to financial activity or “social links” could be a red flag for a scam.
“It’s like playing six degrees of Kevin Bacon, but in a much larger scale,” says Dave Porter, director of the Washington D.C. office of software company Detica.
He said in most banks and insurance companies such a task is done manually with teams of two to 20 personnel sifting through daily transactions and deciding which ones could be questionable and then conducting further investigation. The work could take anywhere from two days to several weeks.
The data mining power of SAS’s Fraud Management BI tool is able to cull information from numerous disparate data source such as checking, savings, credit card, loans and other accounts and sources. Detica’s NetReveal software then sifts through this data to find “hard” (apparent) and “soft” (unobvious) links.
Using metrics and modeling features, investigators can accurately score every transaction at the point of sale in order to identity possible fraud. The system can investigate about one billion connections and 50 million networks, according to Porter.
He said criminals know sheer volume of large number of data that companies have to handle each day makes their lives easier. So-called “dirty data” also makes it harder for investigators to carry out their jobs; sometimes, human nature takes over and probers quit searches when they get tired.
Teams need not wait for days to get results; the automated system can be set to produce reports on schedule.
Investigating social networks can also be used to improve insight into customer preferences and behaviour, according to Ernis Sfakiyanudis, president and CEO eTelemetry Inc., an Annapolis, Md.-based technology company.
The company provides clients with a network appliances such as their Metron and Locate models that can be deployed to help marketers analyze e-mail, Voice over Internet Protocol (VoIP) and SMS traffic as well as other Web surfing behaviour.
“We use business intelligence to establish groups of people based on their closeness and other attributes they share,” said Sfakiyanudis.
eTelemetry enables companies to determine who is using the Internet, what sites are being visited and correlate other data such as surf-time, chat-time and bandwidth use.
Aside from helping companies determine for instance what demographics their products would attract, eTelemetry units can also be used for resource management and risk factor mitigation, said Alan Schunemann, chief technology officer for eTelemtry.
Deployed in an office setting, the appliance can help determine the amount of bandwidth that personnel are using as well as which sites they are visiting.
“Such a procedure can be conducted by IT personnel but it could take away from productivity. eTelemetry can do this task automatically in real time,” Schunemann said.