Non‑parametric method that classifies or regresses based on the majority of nearest samples.