Many merchants and banks leverage rules engines as part of their system to help make fraud prevention less manual. Fraudsters, however, adapt quickly when certain avenues of attack are closed to them and adjust their tactics. Unless merchants regularly update the rules underlying the system, it becomes simple for fraudsters to get around those checks. As the number of rules multiplies, it becomes more difficult to maintain them. But, a Mountain View, Calif.-based technology provider yesterday launched a rules engine that leverages machine learning to do just that.
As part of its overall antifraud technology solution, DataVisor has added what it is call its Automated Rules Engine. The company said its system analyzes rules a company has in place and data that bears on those rules to automatically create new rules and retire old ones when necessary, monitor the effectiveness of rules that are in place and back test manually created rules against historical data.
“While rules certainly have their place within the detection ecosystem, they can be difficult to maintain and aren’t always reliable,” said Yinglian Xie, CEO and co-founder of DataVisor. “Online criminals are quick to change their attack techniques and patterns, and rules quickly become obsolete, placing a huge burden on internal fraud and AML teams to update and tune them constantly. [This solution] not only saves time on creating, testing and deprecating rules, but makes them stronger and more accurate by strengthening them with Unsupervised Machine Learning. It’s a new and strong weapon for our customers to use in the fight against online crime.”
- Identifying and Preventing False Positives
- Identifying and Preventing ATOs and New Account Creation Fraud
- Managing a Successful Fraud Team