EasyDeepLearn

What is stability selection?

hard

Answer

  • Run your feature-selection method (usually Lasso) on many random subsamples of the data.
  • Record how often each feature is selected.
  • Keep only features selected in ≥ πthr\pi_{\mathrm{thr}} (typically 60-90%) of subsamples.
  • Rationale: any single Lasso fit is unstable — features come and go with the sample.
  • Stability selection isolates features that are robustly informative across data resamples, with theoretical false-discovery-rate guarantees.
Check yourself — multiple choice
  • Select features by variance
  • Repeated Lasso on subsamples; keep features selected often — theoretically bounded FDR
  • Only for classification
  • It's the same as bagging

Stability selection: repeated Lasso + keep features selected in ≥ threshold fraction of runs.

#feature-selection#regularization

Practise Supervised Learning

215 interview questions in this topic.

Related questions