1. Load the Training dataset and the data miner model-building node in workspace.

Here the data miner node "Stepwise GDA ANCOVA with Deployment" is used as example. 

Be reminded to put a "Select Variable" node after Training dataset to specify the dependent and predictor variables for the model-building analysis.

2. Double click on the data-miner node to have the "Edit the parameter" dialog.

Under "Deployment" tab, select "Generate PMML code" to "True". Click OK. This is to request the data miner node to output the PMML code for the prediction model. 

3. Run the workspace or the data miner node. 

A reporting document appears. The PMML code for the prediction model is stored in the reporting document. 

Sometimes, the PMML code will appear as a seperate node in workspace, depending on the report routing options set for workspace.

4. Open the reporting document, find the file "GDA PMML script" in the file list on the left.

Click on "Home | Save | Save Item(s) As", and save the PMML code to local drive.

5. Go to Statistica "Data Mining" menu tab, Click on "Rapid Deployment" to have the rapid deployment node added to the workspace

6. Link the Testing data with "Rapid deployment" node.

Be reminded to have the "Select Variable" node after "Testing Data" to specify the dependent and predictor variables for the analysis. 

7. Double click on the "Rapid deployment" node, in the "Edit parameters" dialog, under "General" | "Select PMML models for analysis" option, browse for the PMML code saved in step 4. Click OK.

8. Run the "Rapid Deployment" node.  This "Rapid deployment" node will apply the PMML code on the "Testing dataset" and make predictions. 

The performance of the PMML code on Testing dataset are stored in the "Reporting document" node, and the predictions for the Testing dataset is shown in the downstream spreadsheet node with name "Rapid deployment". You can rename the spreadsheet.