Workspace Node: Azure Machine Learning: Quick Tab and Azure Settings Tab

The Azure Machine Learning workspace node can be accessed from the Feature Finder, the ribbon bar, or the Node Browser. The Specifications - Quick tab is displayed by default.

  • For batch scoring, the size of the data can be large. Azure storage is used between Statistica and the scoring model in Azure.

    Data flow: Statistica → Azure Storage → Azure ML scoring flow → Statistica .

  • For live scoring, the size of the input data is assumed to be small, data transfer size negligible. In this case, the data is supplied directly to the scoring model in Azure ML, bypassing Azure storage.

    Data flow: Statistica → Azure ML scoring flow → Statistica .

Specifications - Quick Tab

Element Name Description
Variables Click the Variables button to display a variable selection dialog box.

Specifications - Azure Settings Tab

Element Name Description
Base URL Enter the URL for the Azure ML base.
API Key Enter the API key.
Web service input This option is only available when the Enable batch scoring check box is not selected. This parameter must match the web service input name that is defined in the Azure ML model on the Microsoft Azure platform.
Web service output This option is only available when the Enable batch scoring check box is not selected. This parameter must match the web service output name that is defined in the Azure ML model on the Microsoft Azure platform.
Enable batch scoring Clear this check box when executing live scoring. Select this check box for batch scoring. When the Enable batch scoring check box is selected, the following four options are displayed:
  • Storage Account Name: Enter the name of the storage account.
  • Storage API key: Enter the storage API key.
  • Storage Container Name: Enter the storage container name.
    • Time out (in minutes): Specify an amount of time, in minutes, for the program to time out. When the scoring request is sent to Azure for scoring, the program will wait this amount of time to get the response back. If the program does not get a response within this period, and error message is displayed and no results are produced.

Processing

Element Name Description
Batch scoring

Upon execution of the node:

Data is sent to Azure Storage.

A scoring flow is executed in Azure.

Results are extracted back to Statistica .

Results are mapped to column names in results spreadsheet.

Live scoring (single case)

Upon execution of the node:

Data is sent directly to Azure ML scoring flow and executed.

Results are extracted results back to Statistica .

Results are mapped to column names in results spreadsheet.