Training a machine learning model

In BMC Helix Edge, training Machine Learning models is essential for the following reasons:

• The Machine Learning model automatically learns the patterns and relationships in the data for optimized decisions.
• A well-trained model makes highly accurate predictions, thereby improving the overall effectiveness of the IoT system.
• A trained model reduces manual intervention by automating complex decision-making processes, leading to more efficient use of resources.
• The model adapts to changing conditions, making the system more resilient and dynamic.

AI and Machine Learning Foundation with BMC Helix Edge

BMC Helix Edge uses the AI and Machine Learning Foundation services to manage algorithms and configurations and to train and manage Machine Learning models. The AI and Machine Learning Foundation services enhance the OT monitoring capabilities of BMC Helix Edge by identifying anomalous patterns through AI-driven solutions.

To train a machine-learning model

You train Machine Learning models to learn patterns and relationships from data. The Machine Learning models use these patterns to predict or decide on new, unseen data. After the Machine Learning models are trained to learn from data and adapt to new information, they assist you in automating tasks, gaining insights, improving decision-making, and optimizing operations across different domains.

1. On the BMC Helix Edge page, click Machine Learning.
2. On the Machine Learning page, identify the specific model in the list of trained models.
3. Clicknext to the model to expand the model details.
4. In the expanded Model panel, click Train a Model.
   The label shows as Manage model training if you have an already trained model. 

5. In the Train a Model panel, enter a unique training ID to create and train the new model and click Submit.
   BMC Helix Edge submits the job to the AI foundation for model training.
   You can monitor the progress of the activity in the console with the following details:

   Column name	Description
   Training ID name	Indicates the unique identifier for the training session.
   Training status	Indicates the status of the training process:  New: Indicates that the Machine Learning model has been newly created but has not yet undergone training. The model is in its initial state and has not started learning from the provided data. TrainingInProgress: Indicates that the Machine Learning model is now undergoing the training process. The model is actively learning patterns and relationships from the provided training data. TrainingCompleted: Indicates that the training process for the Machine Learning model has been completed. The model has finished learning from the training data and is ready for use in making predictions or analyses. TrainingFailed: Indicates that the training process for the Machine Learning model has failed. The model needs investigation for errors and corrected accordingly.
   Training date	Specifies the date on which the training was initiated.
   Training time	Provides the estimated time to train the model and the actual training time after training is complete.
   Version	Displays the versions related to the trained model after completing the training.
   Actions	Offers the following options: Retrain: When you select this option, BMC Helix Edge initiates the model training process with the existing model as a starting point. It uses new data or configurations to update and optimize the model's performance. Delete: When you select this option, BMC Helix Edge removes all associated data, configurations, and information related to the trained model from the system. Delete a trained model if the model is no longer relevant, consumes unnecessary resources, or needs to be replaced with a different model.

6. After successfully training the model, BMC Helix Edge displays the trained model instance in the console with the following details:

   Column name	Description
   Version	Displays the versions related to the trained model.
   Training ID	Displays the unique training identifier.
   Node Deployments	Provides information about node deployments associated with the trained model. This node is the BMC Helix node instance where you deploy this model.
   Operational Status	Shows the operational status of model deployment on the BMC Helix Edge node.
   Actions	Offers the following options: Event generation settings: You can configure how the Machine Learning model generates events or predictions based on incoming data. You define thresholds, rules, or conditions that trigger specific events when the model's predictions meet certain criteria. To learn more, see Configuring Event generation settings.  Deploy: Deployment refers to making a trained machine-learning model available in a production environment. After you train the model, you must deploy the trained machine-learning model to process real-time data and provide predictions or classifications. With this option, you can integrate the model into the existing infrastructure and make it accessible to applications or services that require its functionality. To learn more, see Deploying the trained Machine Learning model.

Example: Video example of how to train a Machine Learning model already created.

To retrain a Machine Learning model

Retraining trained Machine Learning models contributes to your IoT deployment's ongoing success and improvement. It supports historical analysis, provides redundancy, ensures compliance, aids in model improvement, assists in testing, and fosters research and development efforts.  

Use the following procedure to repeat if the initial training process encounters errors and failures: 

1. On the BMC Helix Edge page, select Machine Learning.
2. On the Machine Learning page, identify the specific model in the list of trained models.
3. Clicknext to the model to expand the model details.
4. In the expanded Model panel, click Manage model training.
5. In the Manage model training panel, determine a training ID from the list and select Action > Retrain.
   

1. To check on the retraining progress, click in the top-left corner of the console.

Where to go from here

Deploying-the-trained-Machine-Learning-model