Adding an Anomaly Detection Channel for Model Monitoring

You can add an anomaly detection channel to a ASCMO-DYNAMIC project and train a Dynamic Gaussian Mixture Model for this channel.

This Dynamic Gaussian Mixture Model works on the selected inputs. It is not connected to an RNN and does not use the internal states of an RNN.

To add an anomaly detection channel

1. Select In/Outputs > Add Output for Anomaly Detection.

2. Open the model configuration for the anomaly detection channel.

3. Select Dynamic Gaussian Mixture Model as model type.

4. Configure the model parameters.

5. Start the model training.

6. The anomaly detection channel is trained with a Dynamic Gaussian Mixture Model.

See also  

Model Configurations: Dynamic Gaussian Mixture Model