February 27, 2013
Hacking, fraud and other clandestine online activities have been making headlines in recent weeks, giving rise to concerns that law enforcement agencies are losing the war against cybercriminals. But just how serious a threat to the public is cyber crime, and could data science hold the key to reversing the trend?
RSA, the cyber security arm of US big data firm EMC, specialises in the use of advanced analytics and machine learning to predict and prevent online fraud. Its Anti-Fraud Command Centre (AFCC) has identified and terminated 500,000 such attacks in its eight year existence, half of which came in 2012 alone.
This increasing detection rate is due in no small part to its rapid adoption of machine learning techniques. Five years ago RSA’s Israeli operation undertook a step change, moving away from inflexible rule-based fraud detection systems, in favour of a self-improving approach using data science underpinned by Bayesian inferencing.