04_​Partial_​Dependence_​Pre-processing

0 ×

This is an example for visualizing a partial dependence plot and an ICE curves plot in KNIME.
AutoML (Regression) Component was used to select the best model for the given data, but any model and its set of Learner and Predictor nodes can be used.

- Read the test and train dataset
- Select 100 rows from test dataset to be explained by Partial Dependence/ICE Plots
- Create the samples via the apposite shared component Partial Dependence Pre-processing
- Score the samples using the Workflow Executor and AutoML (Regression) Component.
- Visualize the Partial Dependence/ICE Plot and customize it directly in the View (Execute > Right Click > Open View)
- Apply and Close to save the custom in-view settings

Nodes

• Component Input60 ×
• Component Output60 ×
• Column Filter50 ×
• Column Rename48 ×
• Constant Value Column43 ×
• Show all 130 nodes

Extensions

• FeatureKNIME Base nodes
• FeatureKNIME Data Generation
• FeatureKNIME Deep Learning - Keras Integration
• FeatureKNIME Ensemble Learning Wrappers
• FeatureKNIME H2O Machine Learning Integration
• FeatureKNIME Integrated Deployment
• FeatureKNIME JavaScript Views
• FeatureKNIME Javasnippet
• FeatureKNIME Machine Learning Interpretability Extension
• FeatureKNIME Math Expression (JEP)
• FeatureKNIME Optimization extension
• FeatureKNIME PMML Preprocessing Applier Nodes
• FeatureKNIME Quick Forms
• FeatureKNIME XGBoost Integration

Links

• Christoph Molnar - Interpretable Machine Learning - A Guide for Making Black Box Models Explainable - 5.1 Partial Dependence Plot (PDP)

  christophm.github.io/interpretable-ml-book/pdp.html
• Verified Components

  www.knime.com/verified-components
• Big Mart Sales Prediction Datasets

  www.kaggle.com/datasets/shivan118/big-mart-sales-prediction-datasets