Icon

02_​XAI_​Credit_​Scoring_​Example

Explaining the Credit Scoring Model with Model Interpretability Components

This workflow demonstrates the usage of the verified components developed to interpret machine learning models.

In the example, the Credit Scoring data set is partitioned to training and test samples. Then, the black box model (Neural Network) is trained on the standardly pre-processed training data using the AutoML component. The Workflow Object capturing the pre-processing and the model is provided as one of the inputs for the model interpretability components.

The components allow for both global (Global Surrogates, Permutation Feature Importance, Partial Dependence Plot) and local (Counterfactual Explanations, Local Surrogates, ICE, SHAP) explainability techniques.

URL: KNIME Integrated Deployment - KNIME.com https://www.knime.com/integrated-deployment
URL: Molnar, Christoph. "Interpretable machine learning. A Guide for Making Black Box Models Explainable", 2019. https://christophm.github.io/interpretable-ml-book/
URL: Give Me Some Credit - Kaggle data set https://www.kaggle.com/c/GiveMeSomeCredit/

Data access
Model explanations
This workflow demonstrates the usage of the verified components developed to interpret machine learning models on the example of the credit scoring model.
Data pre-processing, black-box model training, predicting
Select a group of applicants
Scatter Plot Selection
Select applicant wrongly rejected by the model (false negative)
Scatter Plot Selection
Standard pre-processing & black-box model training Neural Network
AutoML
XAI View
Local Explanation View
Global Feature Importance
Undersampling of the frequent "creditworthy" class
Equal Size Sampling
Predictor
Row Sampling
Credit scoring data
CSV Reader
Top: train set Bottom: test set
Partitioning

Nodes

Extensions

Links