Anomaly Detection. Time Series AR Training Dataviz

This workflow trains an auto-regressive model for anomaly detection:
- Filter the data to training data covering only normal functioning
- Visualize the amplitude values on two selected frequency bands
- Loop over each frequency column at a time
- Train an auto-regressive model using 10 past values as predictors
- Calculate in-sample prediction error statistics
- Save the model and prediction error statistics for deployment