A supervised machine learning model has two phases, training and testing. When these models are learned, validated, and tested, the test and train data points are normally presumed to have the same distribution. In the real world, however, the training and test datasets rarely follow the same distribution.
2021-05-22