As a first pass definition it does well to explain the concept. Even if you're interpolating you will need to rank the samples and find the two nearest neighbou

As a first pass definition it does well to explain the concept. Even if you're interpolating you will need to rank the samples and find the two nearest neighbours to interpolate between. It serves to distance it from the moment-based statistics like mean and variance at least.