Title :
Parallel algorithms for hierarchical clustering and cluster validity
Author_Institution :
Dept. of Comput. Sci., Alberta Univ., Edmonton, Alta., Canada
fDate :
11/1/1990 12:00:00 AM
Abstract :
Parallel algorithms on SIMD (single-instruction stream multiple-data stream) machines for hierarchical clustering and cluster validity computation are proposed. The machine model uses a parallel memory system and an alignment network to facilitate parallel access to both pattern matrix and proximity matrix. For a problem with N patterns, the number of memory accesses is reduced from O(N 3) on a sequential machine to O(N2) on an SIMD machine with N PEs
Keywords :
computational complexity; computerised pattern recognition; parallel processing; SIMD machines; alignment network; cluster validity; computational complexity; hierarchical clustering; parallel algorithms; parallel memory system; pattern matrix; proximity matrix; Application software; Clustering algorithms; Computational modeling; Concurrent computing; Data analysis; Euclidean distance; Hypercubes; Parallel algorithms; Partitioning algorithms; Pattern analysis;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
