简介:Withtheexpansionofthegeneexpressionprofiledatabase,inthecaseofaslittleaspossibletoloseinformationortoretainthemostcriticalinformation,geneextractionhasbecomeamaindirectionforthescholars.Thispaperexcludes1561irrelevantgenesthroughthedefinitionofweighteddistancefirstly,andthenremoves252redundantgenesbyPearson'scorrelationcoefficient.Finallybycomparingthetwomethods,stepwiseregressionafterclusteringandonlystepwiseanalysis,weobtainthebestcombinationof8genes.