Research map / Learning theory and algorithms / Prediction and decision losses
Globally-Consistent Rule-Based Summary-Explanations for Machine Learning Models: Application to Credit-Risk Evaluation
Cynthia Rudin, Yaron Shaposhnik
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use BinMaxSupport instead of BinMinSetCover | summary-explanations · FICO HELOC binary | support | Improved |
| use BinMaxSupport instead of BinMinSetCover | summary-explanations · Census 1990 | out-of-sample consistency | Hurt |
| add one feature to the rule (ablation) | BinMaxSupport · FICO HELOC binary | average support: roughly +300 observations | Improved |
| increase explanation train-set size (ablation) | summary-explanations · Census 1990 | test inconsistencies: under 70 -> under 10 (out of 50,000) | Improved |
| increase explanation train-set size (ablation) | summary-explanations · Census 1990 | running time | Hurt |
| increase explanation train-set size (ablation) | summary-explanations · Census 1990 | sparsity | Hurt |
| increase explanation train-set size (ablation) | summary-explanations · Census 1990 | support | Improved |
| use summary-explanations instead of Anchors | summary-explanations · Census 1990 | inconsistencies | Improved |
| use BinMaxSupport instead of Anchors | summary-explanations · Census 1990 | support size | Improved |
| use BinMinSetCover instead of Anchors | summary-explanations · Census 1990 | support size | Hurt |
| use summary-explanations instead of Anchors | summary-explanations · Census 1990 | sparsity and runtime | Improved |
| change global model (KNN, LR, SVM, CART, RF, AdaBoost) | summary-explanations · FICO HELOC | rule sparsity and support | No clear effect |
Closest papers by meaning
- Local vs. Global Interpretability: A Computational Complexity Perspective (ICML 2024 · similarity 0.85)
- Feature Responsiveness Scores: Model-Agnostic Explanations for Recourse (ICLR 2025 · similarity 0.85)
- Prediction without Preclusion: Recourse Verification with Reachable Sets (ICLR 2024 · similarity 0.84)
- Additive Models Explained: A Computational Complexity Approach (NeurIPS 2025 · similarity 0.83)
- Probabilistic Explanations for Linear Models (AAAI 2025 · similarity 0.83)
- Auditing Local Explanations is Hard (NeurIPS 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
