#Computers Like K-Means, but it selects the prototypes to be the medoid rather than the average of points in the cluster. Less sensitive outliers