User Guide

• 1. Introduction
  • 1.1. Introduction
  • 1.2. Toolbox Design
  • 1.3. Interpretable Models
  • 1.4. Diagnostic Suite
  • 1.5. Future Plan
• 2. Data Pipeline
  • 2.1. Data Load
  • 2.2. Data Summary
  • 2.3. Data Preparation
  • 2.4. Exploratory Analysis
  • 2.5. Data Quality (Integrity Check)
  • 2.6. Data Quality (Outlier Detection)
  • 2.7. Data Quality (Drift Test)
  • 2.8. Feature Selection
• 3. Model Train and Tune
  • 3.1. Train and Register Sklearn Style Model
  • 3.2. Register H2O Models
  • 3.3. Register Arbitrary Models
  • 3.4. Hyperparameter Optimization (Model Tune)
• 4. Post-hoc Explainability
  • 4.1. PFI (Permutation Feature Importance)
  • 4.2. PDP (Partial Dependence Plot)
  • 4.3. Hstats (Friedman’s H-statistic)
  • 4.4. ICE (Individual Conditional Expectation)
  • 4.5. ALE (Accumulated Local Effects)
  • 4.6. LIME (Local Interpretable Model-Agnostic Explanation)
  • 4.7. SHAP (SHapley Additive exPlanations)
• 5. Interpretable Models
  • 5.1. Generalized Linear Models
  • 5.2. Generalized Additive Model
  • 5.3. Decision Tree
  • 5.4. Fast Interpretable Greedy-tree Sums
  • 5.5. XGBoost Depth 1
  • 5.6. XGBoost Depth 2
  • 5.7. Explainable Boosting Machines
  • 5.8. GAMI-Net
  • 5.9. ReLU Neural Network
• 6. Diagnostic Suite
  • 6.1. Accuracy
  • 6.2. WeakSpot
  • 6.3. Overfit
  • 6.4. Reliability
  • 6.5. Robustness
  • 6.6. Resilience
  • 6.7. Fairness
  • 6.8. Segmented
  • 6.9. Scored Test
• 7. Model Comparison
  • 7.1. Comparison for Regression
  • 7.2. Comparison for Classification
  • 7.3. Fairness Comparison
• 8. Case Studies
  • 8.1. BikeSharing Data
  • 8.2. CaliforniaHousing Data
  • 8.3. TaiwanCredit Data
  • 8.4. Fairness Simulation Study 1
  • 8.5. Fairness Simulation Study 2