Investigation ofanewGRASP-basedclusteringalgorithmappliedtobiologicaldata

A largeamountofbiologicaldatahasbeenproducedinthelastyears.Importantknowledgecanbe extracted fromthesedatabytheuseofdataanalysistechniques.Clusteringplaysanimportantrolein data analysis,byorganizingsimilarobjectsfromadatasetintomeaningfulgroups.Severalclustering algorithms havebeenproposedintheliterature.However,eachalgorithmhasitsbias,beingmoreade- quate forparticulardatasets.Thispaperpresentsamathematicalformulationtosupportthecreationof consistent clustersforbiologicaldata.Moreover,itshowsaclusteringalgorithmtosolvethisformulation that usesGRASP(GreedyRandomizedAdaptiveSearchProcedure).Wecomparedtheproposedalgorithm with threeknownotheralgorithms.Theproposedalgorithmpresentedthebestclusteringresultscon- firmed statistically.