EasyDeepLearn

Outliers vs drift — how do you distinguish?

medium

Answer

  • Outliers = individual points far from distribution; drift = the distribution itself shifts.
  • Time-based test: (1) if score spikes then returns → outliers or anomalies.
  • (2) if score persistently elevated → drift.
  • Aggregate score over rolling window (24h).
  • Individual outlier detection: Isolation Forest, LOF, autoencoder reconstruction error.
  • Drift detection: aggregated PSI / MMD over batch.
  • Both matter but require different responses.
Check yourself — multiple choice
  • Same thing
  • Outliers = individual far points (spike then recover); drift = distribution persistently shifts; rolling window differentiates + different responses required
  • Random
  • Only outliers

Outlier = point; drift = distribution; rolling window differentiates.

#monitoring#data-quality

Practise MLOps & Data Quality

215 interview questions in this topic.

Related questions