◷ Reading Time: 5 minutes
PMML (Predictive Model Markup Language) is an XML-based language. It contains definitions of predictive models and enables the sharing of those models between applications.
Once you create the PMML file of your predictive model, you can import and use it in FlexRule.
Once the Analytics Extension is installed, PMML can be consumed in decision models. It can be directly used in a Decision Requirement Diagram (DRD), or in a Flow.
The PMML supports the algorithms shown below:
- Regression Model
- Generalized Regression Model
- Naive Bayes Model
- Classification and Regression Tree Model
- Support Vector Machine Model
- Random Forest (Mining Model)
- Neural Networks Model
- Clustering Model
- Gradient Boosting Machine Model (Mining Model)
- Association Rules Model
- K-Nearest Neighbors Model
In FlexRule Designer, go to,
Documents --> PMML
You will see the following window.
Add the PMML file location and a name for the Fact (It will create a Fact Concept file and this will be the name of the fact).
Once the file is imported successfully, you will be able to see the PMML file and its Fact Concept file. It automatically creates the Fact Concept file.
View of PMML from XML File Editor
You can open the imported file on Project Explorer.
Click on the button at the bottom left corner, if you want to open the related Fact Concept file.
Use of PMML in a Decision Requirement Diagram
Once you import a PMML file to your project, you can reference it in a DRD (Decision Requirement Diagram).
1. Create a DRD file in your project.
2. Add an InputData node and set its properties.
- Name: Give a name to the node
- Assembly: Select the concept file from the list
- Type: Name of the Fact in the Fact Concept file
4. Define an output variable to which to copy results.
3. Drag and drop the imported PMML file and set its properties.
- Copy Value To: The output variable to which to copy the result
- Uri: PMML file path
- Parameters: Related parameter
These are the sample projects that use PMML.