The Scoring node performs scoring by utilizing the columns of a data table as input to a trained classifier and adds columns containing predictions and associated probabilities of each class.
An Execution Profile with an active session linked to the workflow
A suitable model generated from the Auto ML node
A Data Frame node - table generated from the import node on which the Scoring has to be run
From the Machine Learning group of the Nodes tabbed page, drag and drop the Scoring node to the workflow canvas.
Connect the output of the model to the Model socket and a new data table to the Table Socket of the Scoring node.
Configure the Scoring node by double-clicking the node or using the Open option provided in the node menu.
Click Save and Run.
Double click the Scoring Result object node to view information, such as Variable Information, Chart, Column metrics, and so on.
The Scoring node configuration tabbed page displays columns from both the Model and Table side by side and the matching columns are auto mapped.
You can manually select and map the appropriate columns from the Model and Tables lists if the names in the scoring table differ from those in the training table.
You can also rename the table name.
The Scoring Result object node is created with the scoring fields added to the original table.
The prediction columns that were not present in the test data set are now added to the Scoring node.