简介:AhybridGA(geneticalgorithm)-basedclustering(HGACLUS)schema,combiningmeritsoftheSimulatedAnnealing,wasdescribedforfindinganoptimalornear-optimalsetofmedoids.Thisschemamaximizedtheclusteringsuccessbyachievinginternalclustercohesionandexternalclusterisolation.TheperformanceofHGACLUSandothermethodswascomparedbyusingsimulateddataandopenmicroarraygene-expressiondatasets.HGACLUSwasgenerallyfoundtobemoreaccurateandrobustthanothermethodsdiscussedinthispaperbytheexactvalidationstrategyandtheexplicitclusternumber.