4. Click "Train new model".
5. Enter a name for the anomaly detection model and an optional description.
6. Select the data which will be considered for the anomaly detection model. Multiple time
ranges can be selected.
7. Save draft and train later or save and start training. The trained anomaly detection model will
be available for selection here after a few minutes:
8. To start the continuous detection of anomalies, click "Run model".
The status indicates the current step of processing the data. Available statuses:
– Draft
– ReadyForTraining, PreparingForTraining, PreparingFailed
– Training
– TrainingFailed, ReadyForDeployment, Deployed
– DeploymentFailed
Note
Editing an anomaly detection model
After an anomaly detection model has been edited, the model must be trained again.
Operating
8.1 Training anomaly detection models
Smart Condition Monitoring with IIoT Sensors
48
Operating Instructions, 08/2021, A5E50573389-AA