Predictive Model rule form
Completing the Predictive model tab

This step is not required when the predictive model instance is generated by the Predictive Analytics Director functionality directly available in Decision Management enabled applications, as the predictive model instances generated in the model export process already include the OXL file.

Model upload

In this section, you upload the file containing the model. The model is represented by the OXL files if you are creating a predictive model based on a PAD model, or the PMML files if you are creating a predictive model based on a third party model.

• Click Upload model to select the appropriate file. When the upload is successful, information about the predictive model is displayed including the file and model details as well as, for PAD models, the model performance measured in terms of Coefficient of Concordance (CoC).
• Click Upload different model to upload a new version of the model.

Note: Spaces are not supported in the file name. Make sure you rename the OXL file to avoid naming conflicts on the Pega 7 Platform. If you are using OXL files generated by the CDM PAD application, take into consideration that not every OXL version contains aggregation information; in this case, use models generated by CDM PAD 6.1.2 or lower, or CDM PAD 6.8 or higher.

The PMML supported models

• Clustering
• GeneralRegressionModel
• MiningModel
• NaiveBayesModel
• NearestNeighborModel
• NeuralNetwork
• RegressionModel
• RuleSetModel
• Scorecard
• SupportVectorMachineModel
• TreeModel

Model details

When you upload a model, this section displays the summary of the model (name, type, and description). Here, you can also click the Download file link to download the OXL or PMML file.

Model XML

This section provides the XML schema preview for the model that you uploaded. You can check the structure of the model and make edits if necessary.

Errors

This section appears when you try to save the uploaded model that contains syntax errors or is not supported. Go to this section to check which line contains the error.

XSD validation and error messages

When you upload a PMML file in the Predictive Model rule and want to save it, the file is parsed and checked for any syntactic errors. The contents of the PMML file is validated against the respective version of the XSD schema specified in the file. For the list of errors, check the following table.

Error code	Description
ATTRIBUTE_NOT_ALLOWED	The attribute is not allowed in the tag.
INVALID_ATTRIBUTE_ID/NAME	The attribute name/ID is incorrect.
INVALID_DECLARATION	The declaration of PMML or any other tag is not correct.
INVALID_ELEMENT_INFO	The specified element values information is not correct.
INVALID_ELEMENT_LOCATION	The specified attribute is not in the correct location of the tag.
INVALID_ID_REFERENCE	The element reference ID is not correct.
INVALID_SCHEMA_SYMBOLS	The specified schema symbols are not valid.
INVALID_TYPE_DECLARATION	The specified type is not valid.
INVALID_TYPE_VALUE	The data type value present is not correct.
INVALID_XSI_TYPE	The xsi schema is not correct.
VALUE_NOT_IN_CONSTRAINT	The value specified for an attribute is not in the constraint for that attribute.

Open topic with navigation