Why is k-NN called a lazy learner?

A. It stores training examples and postpones most computation until prediction B. It learns only from missing values C. It refuses to use labels D. It always uses fewer than two observations