EasyDeepLearn

Semi-supervised vs unsupervised anomaly detection.

medium

Answer

  • Unsupervised: no labels; assumes normals dominate.
  • Semi-supervised: only normal data labeled — train on clean normals, detect deviations (one-class SVM, deep SVDD, AE).
  • Supervised: rare — a few labeled anomalies (PU learning, imbalanced classification).
  • Rule: whenever you have even 20-50 confirmed anomaly examples, do semi-supervised evaluation or PU learning — it's a huge lift over pure unsupervised.
Check yourself — multiple choice
  • Same thing
  • Unsupervised (no labels, assume normals dominate) / semi-supervised (clean normal labeled) / supervised (rare, PU-learning); 20-50 anomaly labels are already a big lift
  • Random
  • Only supervised

AD label regimes: unsup / semi-sup (normal-only) / PU / supervised; leverage any labels.

#anomaly-detection

Practise Unsupervised Learning

214 interview questions in this topic.

Related questions