Working with ML predictions
Once you have deployed your machine learning model, you can use the model to create predictions. These predictions can be used to make more efficient and informed decisions based on your data.
You can create and edit ML deployments and generate predictions in personal or shared spaces. You can also publish ML deployments to managed spaces and generate predictions. Access to ML deployments is controlled through the space. For more information about spaces, see Working in spaces.
ML deployments can be created in personal, shared, and managed spaces. Prediction data generated from an ML deployment can be stored in a personal, shared, or managed space.
Workflow
The following steps are an example of how to work with ML deployments and predictions.
- Deploy your model
Deploy the model you want to use to make predictions.
- Make predictions
Make manual or scheduled predictions on datasets or use the prediction API.
- Get your model approved
Before you can make predictions with the ML deployment, the source model needs to be activated for making predictions. Model approval can be performed by users and administrators with specific permissions.
- Visualize the predictive insights
Load the generated prediction data into an app and create visualizations.
- Explore the data with what-if scenarios
Integrate the prediction API into an app to get real-time predictions. This allows you to try out what-if scenarios by changing feature values and getting predicted outcomes for the new values. The record is passed to the ML deployment via API and a response is received in real time. For example, what would happen to the risk of customer churn if we changed the plan type or increased the base fee?
- Take action
Analyze the predictive insights and scenarios to find out which actions to take. Qlik Application Automation helps you automate the actions and provides specific templates for machine learning use cases. For more information about automations, see Qlik Application Automation.
Requirements and permissions
This section lists the user requirements for working with ML deployments, and the predictions you make with them.
ML deployments
To work with ML deployments, you need:
-
Professional or Full User entitlement
-
View and create ML deployments: Automl Deployment Contributor or Automl Experiment Contributor security role
-
Edit and delete ML deployments: Automl Deployment Contributor security role
-
Required role in the space where the ML deployment is located.
For more information, see:
Predictions
To create, edit, and delete prediction configurations, you need:
-
Professional or Full User entitlement
-
Automl Deployment Contributor security role
-
Required role in the space where the ML deployment is located.
Predictions can run as batch predictions (from a prediction configuration) or real-time predictions. You can use the Qlik AutoML connector to run predictions as well.
To run predictions using any of these methods, you need:
-
Automl Deployment Contributor security role
-
Required role in the space where the ML deployment is located:
-
Shared spaces: Users with Professional or Full User entitlement need the Owner, Can manage, Can edit, or Can consume data role in the space. Users with Analyzer entitlement need the Owner or Can consume data role in the space.
-
Managed spaces: Users with Professional or Full User entitlement need the Owner, Can manage, or Can consume data role in the space. Users with Analyzer entitlement need the Owner or Can consume data role in the space.
-
-
For scheduled predictions configured with the AutoML user interface, there are also requirements for the owner of the prediction configuration. See: Prediction configuration ownership
Predictions generated from the Qlik AutoML user interface are created as datasets. Therefore, the same requirements for working with data sources in Qlik Cloud apply to working with this prediction output (such as using it in a Qlik Sense app). You must have the Private Analytics Content Creator role to create datasets in your personal space.
For more information, see:
Model approval
To activate and deactivate the source deployed model for an ML deployment, you need specific permissions. These permissions are different depending on whether you are performing these actions as a user or administrator. For more information, see: