Implement covariate shifting as a method of spatial validation. Rough idea is to train a classifier to on a mix of training and test data (from a prediction area). The greater the ability to distinguish between the two types, the greater the shift and the less reliable the model comes.
Implement covariate shifting as a method of spatial validation. Rough idea is to train a classifier to on a mix of training and test data (from a prediction area). The greater the ability to distinguish between the two types, the greater the shift and the less reliable the model comes.