As far as I know, the real concept drift is caused by changes in the decision boundary while virtual drift occurs because of changes in data distribution. Some researchers mention that virtual drift can be denoted as feature change.

Is my information correct?

Does feature drift, where features become or cease to become relevant with time is considered as virtual or real concept drift, or it is a different type?


1 Answer 1


About the data drift types, you can consider these generic types:

  • features drift: given when the distribution of the input features (comparing training datasets VS prediction datasets) change enough (with a defined threshold) to raise an alert
  • target drift: distribution of the label values change when comparing training VS prediction distributions
  • concept drit: when the relation between the input features and target values change; it can arise for instance when the label is redefined

A bit more detail in this validated answer


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.