Tree-Based Machine Learning Methods for Prediction and Variable Selection

Part I       Part II       Part III       Part IV

Hemant Ishwaran and Min Lu
University of Miami

Outline

Part I: Training

  1. Quick start
  2. Data structures allowed
  3. Training (grow) with examples
    (regression, classification, survival)

Part II: Inference and Prediction

  1. Inference (OOB)
  2. Prediction Error
  3. Prediction
  4. Restore
  5. Partial Plots

Part III: Variable Selection

  1. VIMP
  2. Subsampling (Confidence Intervals)
  3. Minimal Depth
  4. VarPro

Part IV: Advanced Examples

  1. Class Imbalanced Data
  2. Competing Risks
  3. Multivariate Forests
  4. Missing data imputation