Detecting Dataset Drift and Non-IID Sampling via k-Nearest Neighbors

Re-assign community