Clustering on a hypercube multicomputer

Squared-error clustering algorithms for single-instruction multiple-data (SIMD) hypercubes are presented. These algorithms are asymptotically faster than previous algorithms and require less memory per processing element. For a clustering problem with N patterns, M features per pattern, and K clusters, the algorithms complete it in O(K+log NM) steps on NM processor hypercubes. This is optimal up to a constant factor. Experimental results from a commercially available multiple-instruction multiple-data (MIMD) medium-grain hypercube show that the clustering problem can be solved efficiently by the machines