Interpretable Models

• Generalized Linear Models
  • GLM in MoDeVa
  • Global Interpretation
  • Local Interpretation
  • Examples
• Decision Tree
  • Decision Tree in MoDeVa
  • Global Interpretation
  • Local Interpretation
  • Examples
• Gradient Boosted Decision Trees
  • Mathematical Formulation
  • GBDT in MoDeVa
  • Interpretability Through Functional ANOVA
  • Global Interpretation
  • Local Interpretation
  • Monotonicity Constraint in GBDT
  • Examples
• Linear Tree and Gradient Boosted Linear Trees
  • Linear Tree
  • Linear Tree in MoDeVa
  • Gradient Boosted Linear Tree (GBLT)
  • GBLT in MoDeVa
  • Global Interpretation
  • Local Interpretation
  • Examples
• Neural Tree
  • Model architecture
  • Neural Tree in MoDeVa
  • Functional ANOVA Representation
  • Global Interpretation
  • Local Interpretation
  • Monotonicity Constraints in Neural Tree
  • Examples
• GAMI-Net
  • GAMI-Net in MoDeVa
  • Functional ANOVA Representation
  • Global Interpretation
  • Local Interpretation
  • Monotonicity Constraints in GAMI-Net
  • Examples
• ReLU Neural Network
  • Model architecture
  • Local Linear Models (LLM)
  • ReLU DNN in MoDeVa
  • Model Interpretation
  • Local Interpretation
  • Examples
• Mixture of Experts (MoE)
  • Model Architecture
  • MoE in MoDeVa
  • Interpretation: Functional ANOVA Representation
  • Global Interpretation
  • Local Interpretation
  • Examples
• DirectRS
  • Overview
  • Model Architecture
  • Implementation in MoDeVa
  • Interpretability Features
  • FANOVA Comparison
  • Key Classes
• ICL-MoE
  • Overview
  • Model Architecture
  • Implementation in MoDeVa
  • Interpretability Features
  • FANOVA Comparison
  • Key Classes
  • Parameters
• GBDT Leaf Kernel
  • Overview
  • Prediction Heads
  • Implementation in MoDeVa
  • Classification
  • Local Evidence (TKLE)
  • Head Nomination
  • Global Importance and Weakness
  • In-Context Prediction
  • Key Classes
• FuseKernel
  • Overview
  • Model Architecture
  • Implementation in MoDeVa
  • Classification
  • Inherent Interpretation (FANOVA)
  • Channel Decomposition
  • Inherent Predictive Uncertainty
  • Diagnose and Repair
  • Key Classes