There are several options for generating predictions in AutoML. In this article, we will outline steps for doing this in a Qlik Sense App using the Qlik_AutoML (non Legacy) data connection. An advantage for using this method, is you can continually refresh the apply or prediction dataset and run the predictions script. This keeps all steps within Qlik Cloud, and is another option to running it manually, on a scheduled basis, or externally with the API.
In this example, I am working with the Iris dataset.
1. Upload iris.csv to Qlik Cloud.
2. Click Add New -> ML experiment and fill out the information
3. Run the experiment with 'variety' as the target
4. Select the champion model and click Deploy
5. Go back to Analytic Services and open a new Qlik Sense App.
Select the prediction data and then click load into App.
*NOTE* in this example, we are using the same dataset for Training as we are for the Prediction dataset. This is not best practice. When selecting a table to base your App on in this step (#5), this is the Apply or Prediction table.
6. Open up Data Load Editor and add a new section called automl_pred
7. Within this section, navigate to the right hand side of the page and click 'Create New Connection'
Select 'Qlik AutoML' under the Analytics Sources section. Note you must be licensed to use this for this source to appear.
Select the ML Deployment we created in #4
Select a name for the Returned table. In this case we include the Apply dataset in the table. Click Create.
8. You will now see a Qlik_AutoML connection icon in your pane. Click on the data icon with the pencil icon.
For the resident table, select 'iris' which is the Apply or predictions dataset we loaded in #5.
Then select 'Iris_predictions_out' under the tables section.
Click on insert script and this will copy the syntax to your data load editor section.
9. Optional: If you would like to save the predictions dataset in Qlik Catalog as a QVD, add the following syntax.
Store Iris_predictions_out into [lib://DataFiles/iris_predictions_output.qvd];
10. Load your data and explore the results in Qlik Sense
The information in this article is provided as-is and to be used at own discretion. Depending on tool(s) used, customization(s), and/or other factors ongoing support on the solution below may not be provided by Qlik Support.
Generating Predictions externally with the API