Interpretability View - Preview (PMML)

0 ×

--------------- !!! DISCLAIMER !!! ---------------

This Component is a not-verified Beta version shared in Fall 2019 on my personal space for demo purposes.

Official Components verified by the KNIME team are available via the Example Server at:

hub.knime.com/knime/spaces/Examples/latest/00_Components

--------------- !!! \\\ END /// !!! --------------

This component computes Shapley Values, Partial Dependence, Individual Contional Expectation Curves and Surrogate Decision Tree View for a model of your choice produced by the "AutoML" Component.

Options

Maximum Sample Size to Explain

Explaining predictions takes some computation time. Decrease this number if you want to compute the dashboard in less time.

Name of your model

Type in here the name of your model.

Predicted Class Category to Explain

To explain prediction of a classification model you have to select the target class category. In the case of binary classification we suggests to use the positive class.

Nodes

• Column Filter9 ×
• Column Rename (Regex)8 ×
• String Manipulation (Variable)8 ×
• RowID6 ×
• Joiner5 ×
• Show all 55 nodes

Extensions

• FeatureKNIME Base nodes
• FeatureKNIME Data Generation
• FeatureKNIME JavaScript Views
• FeatureKNIME Javasnippet
• FeatureKNIME JPMML Integration
• Show all 8 modules

Links

• No links available