With worldwide losses to fraudulent activity now as high as $200 billion a year, 3 data science in general and machine learning, in particular, are proving adept at detecting and preventing financial services fraud—and, in turn, improving security for customers and staff. 4 While more traditional rule-based detection systems can spot obviously fraudulent activity, they subject customers to multiple, tortuous steps and involve lots of manual adjustment to different scenarios. Above all, they can easilymiss subtle or disguised behavior that may also indicate fraud. On all these fronts, machine learning makes a better alternative as a fraud detective. It finds hidden or implicit correlations in data, reduces verification measures, and automatically detects potential fraud in real time. 5 While rule-based systems are still prevalent, leading financial insti- tutions already use data science technology to combat fraud. For example, MasterCard has trained up and integrated machine learning and AI tools to track and process a range of variables across credit card transactions—from transaction size, location, and time to device and purchase data. By providing a real-time judgment on whether a transaction is fraudulent, MasterCard is helping reduce the number of false declines in merchant payments.
FRAUD DETECT ION
1
While rule-based systems are still prevalent, leading financial institutions already use data science technology to combat fraud.
© 2019 PROVENIR ALL RIGHTS RESERVED
Made with FlippingBook Digital Publishing Software