How does MiniBatchKMeans differ from KMeans in Scikit-Learn?

MiniBatchKMeans updates centroids using small random batches instead of full dataset iterations.

Standard KMeans computes full dataset distances each iteration, making it expensive for large datasets. MiniBatchKMeans approximates centroids using mini-batches, significantly improving speed at the cost of slightly reduced accuracy. It is particularly useful in streaming or large-scale clustering scenarios.