EasyDeepLearn

Covariate drift vs concept drift vs label drift — the differences.

medium

Answer

  • Covariate drift: P(X) changes, P(Y    X)P(Y\; \mid \;X) same — model still valid, just used on different inputs (retrain if severe).
  • Concept drift: P(Y    X)P(Y\; \mid \;X) changes — model relationship broken, must retrain.
  • Label drift: P(Y) shifts (imbalance changes) — matters most for calibration.
  • Rule: monitor all three.
  • Concept drift is the hardest to detect without labels (needs delayed ground truth) → proxy via prediction confidence drift.
Check yourself — multiple choice
  • Same thing
  • Covariate: P(X) shifts, model still valid; Concept: P(YX)P(Y \mid X) shifts, retrain needed; Label: P(Y) shifts, calibration; monitor all three
  • Random
  • Only one

Drift types: covariate (P(X)) / concept (P(YX))  /  label(P(Y \mid X))\; / \;\mathrm{label} (P(Y)).

#anomaly-detection#applications

Practise Unsupervised Learning

214 interview questions in this topic.

Related questions