Data: Today Preemptive, Tomorrow Predictive
COLLECTING DATA is all well and good, but the real question is how multiple data points can be intelligently reviewed to help security personnel make better decisions in real time. That's where security analytics comes in.
An overview of security analytics is presented in a new report jointly authored by Mark Sauter, chief operating officer and senior analyst at the Chesapeake Innovation Center, and Scott Greiper, a principal at the financial services firm of C.E. Unterberg, Towin.
Security analytics involves deriving value from security-related data ranging from Suspicious Activity Reports filed by banks to the voice prints or text of phoned-in bomb threats.
Today's uses of these data focus on detecting real-time misbehavior or anomalies and identifying evidence of past criminal behavior: consider motion-tracking CCTV software that detects when a vehicle is moving in the wrong direction or at excessive speed, for example. The next generation will go beyond that and attempt to predict terrorism, fraud, and other undesirable behavior, says Sauter.
Predictive data analytics can be applied in areas ranging from fraud prevention to access control.
One company that provides entry-detection services for courthouses and sensitive government facilities, for example, has started using a "sensor package" at the entrances to buildings. Instead of just using x-ray machines and walk-through metal detectors, the company integrates x-ray images and metaldetector alarms with video, then performs data mining to look for patterns, such as certain clothing or bags that might be associated with a security issue or threat. Software looks for these signs of trouble and signals security personnel before the threat passes the checkpoint.
"Pattern detection over time lets you identify problems or correlate [patterns] to incidents," says Sauter. "It's not just looking back at old cases, but identifying predictors," so that security can decide in real time whether to take an action. Alternatively, similar analytic software might help a company decide whether to accept an insurance claim.