Topic: Subset and Stepwise Selection

Forced, Excluded, and Grouped Predictors

Forward Selection

Entry Rules, Removal Rules, and Stopping

Exhaustive Best-Subset Search

Defining Candidate Model Sets

Correlated Predictors and Path Instability

Computational Limits of Exhaustive Search

Confidence Intervals After Model Selection

Comparing Stability Across Selection Procedures

Choosing Model Size Across Subsets