Case Study: PayPal boosts fraud detection accuracy and speeds model development with H2O Driverless AI

PayPal raises fraud model accuracy to 0.947 and speeds development 6X with H2O.ai

PayPal, a global leader in digital payments, faced the significant challenge of detecting sophisticated and evolving fraud in real time to prevent direct financial losses. To enhance its fraud detection capabilities, PayPal employed H2O.ai's H2O Driverless AI to improve the accuracy and speed of its modeling efforts.

H2O.ai's solution enabled PayPal to integrate expert-engineered features with new features automatically extracted from a massive graph network. Using H2O Driverless AI, the vendor helped PayPal increase its model accuracy from 0.89 to 0.947, a nearly 6% improvement. Furthermore, the solution cut model development time by six times when running on specialized GPU hardware, allowing the team to train models faster and more effectively.

PayPal