Sensor Anomaly Detection

Self-directed study in time-series machine learning, run alongside coursework. Benchmarks classical statistical baselines (rolling z-score, IQR) against learned approaches for anomaly detection in sensor streams. Designed as a reusable framework adaptable to different sensor types and domains.