Medical Device Performance Anomaly Detection System

Construct an advanced anomaly detection system for medical device performance using time-series analysis and machine learning. Develop a solution that can process high-frequency sensor data from medical devices, identifying potential failures or performance degradations before they become critical. Implement a multi-stage approach including data preprocessing, feature extraction, and anomaly detection using ensemble machine learning techniques. The system must have near-zero false positive rates and provide interpretable insights for medical technicians.