Apply Models to New Data

tell me more icon Tell me more

You can apply all of the models to new data, or you can select a model and apply just that model to new data. In either case, the text column in the new data must be prepared in the same way as the text column was prepared in the data used to build the models. In this example we apply all four models to the new data.

  1. In the Component Palette, expand the Data section. Create a data source node using the view MINING_DATA_TEXT_APPLY_V.

  2. In the Component Palette, expand the Text section. Create an Apply Text node. Name the node PrepareApply.

  3. Connect MINING_DATA_TEXT_APPLY to PrepareApply. Connect PrepareBuild to PrepareApply.
    tell me more icon Tell what I see in the GUI.

  4. Right-click PrepareApply and select Run.

    Apply Text applies the text transformations PrepareBuild to the new data. Apply Text does this by automatically matching (by name) columns of MINING_DATA_TEXT_APPLY_V with the transform columns generated in the PrepareBuild node. In this case, Apply Text transforms the attribute COMMENTS in the same way that it was transformed in PrepareBuild.

  5. Monitor the progress of the build in the Workflow Jobs.

    After execution completes, you can apply the models to the prepared new data.

  6. Open the Evaluate and Apply section of the Components Palette. Drag an Apply node to the workflow; name the node Predict.

  7. Link TextClassification to Predict, the apply node; also link PrepareApply to Predict. If you wish, add CUST_ID as data for the apply; this allows you identify predictions.

  8. To apply the four models to the new data, right-click Predict and select Run.

tell me more icon Tell me more about what I see in the GUI.

After the execution completes, you can examine the predictions.