7.1.1 Partitioning Around Medoids (Pam) Compared to the well - known kmeans algorithm, .Pam has the following features: It operates on the dissimilarity matrix of the given data set or when it is presented with an n ´ p data matrix, the algorithm first computes a dissimilarity matrix. It is more robust, because it minimizes a sum of dissimilarities instead of a sum of squared Euclidean dis