Covariate Shift

Covariate shift refers to a situation where the input data distribution changes between training and testing, even though the conditional label distribution remains the same.

Resources

• https://www.seldon.io/what-is-covariate-shift/