Design a Fraud Detection System

Design a scalable fraud detection system that can analyze millions of transactions in real-time to identify and prevent fraudulent activities. Focus on the architecture for processing high-volume data streams, implementing both rule-based and machine learning models for fraud detection, and strategies for continuously improving detection accuracy. Address challenges like handling false positives, adapting to new fraud patterns, and balancing between stringent security and user experience.