MLOps Academy · Lesson

Univariate vs Multivariate Drift

Catch shifts single features alone would miss.

One Feature at a Time

The simplest way to watch for drift is to check each feature on its own. That is called univariate drift detection, and it is where most teams start. 📊

How Univariate Checks Work

For univariate drift, you compare each feature's current distribution against its training reference, one column at a time, with tests like PSI or the KS statistic.

All lessons in this course

  1. Univariate vs Multivariate Drift
  2. Monitor Performance with Delayed Labels
  3. Tune Windows to Cut False Alarms
  4. Wire Drift Alerts to Retraining
← Back to MLOps Academy