How to perform advanced I-tree modelling from training dataset and deploy the PMML model to predict testing or new dataset in workspace?
 

1. Load the training dataset for model generation and testing or new dataset for prediciton in workspace

2. Go to "Node Browser", under "All" list, find and load the Advanced I-tree modelling analysis node, for example, "Advanced Classification Trees (C&RT)". 

3. Double click on the Advanced I-tree analysis node to specify the dependent and independent variables for model building. 

4. In "C&RT Classification" dialog, Continue to click on "Code Generator" under "Result" tab, tick the types of code that you want to output, for example, PMML, SVB, C++, SAS, and so on. Click OK.

5. Run the workspace or the Advanced I-tree analysis node, 3 new nodes will be generated: Reporting document, PMML model, and Rapid Deployment node. "PMML model" node is generated from the I-tree analysis and is automatically connected with "Rapid deployment" node for new dataset prediction. 

You can view the PMML code of the I-tree model generated by clicking on the "PMML model" node.

You can view and save other model codes to your local drive from the "Reporting Document", like SVB code, C++ code, PMML code, etc.

6. The "PMML model" node stores the PMML model generated from the I-tree analysis and is ready to be deployed to predict new dataset through "Rapid deployment" node. 

Link the new dataset (e.g. TestingDataset) to the "Rapid deployment" node, and run workspace. Prediction of the TestingDataset will be stored in "Reporting Document". 

Click on the "Reporting Document", you can find the prediction result of the TestingDataset.