Machines fail. By creating a time-series prediction model from historical sensor data, you can know when that failure is coming Anomaly detection covers a large number of data analytics use cases.
Dr. James McCaffrey from Microsoft Research presents a complete program that uses the Python language LightGBM system to create a custom autoencoder for data anomaly detection. You can easily adapt ...