K-meansClusteringalgorithm1K-meansClusteringK-meansclusteringisasortofclusteringalgorithmanditisamethodofvectorquantization,originallyfromsignalprocessing,-meansclusteringaimstopartitionnobservationsintokclustersinwhicheachobservationbelongstotheclusterwiththenearestmean,servingasaprototypeofthecluster.--pletesamplesastheinitialcenter.(That'swhatkmeansinK-means)Eachpointinthedatasetisassignedtotheclosedcluster,',,B...H,Jintotwoclusters?4A(1,4)B(2,4)CDEFIGJHRandomlychooseA,BasthecentreandK=<<>>>>>>>>So,weclassifyA,CasaclusterandB,E,D,F,G,H,→B5A(1,4)B(2,4)CDEFIGJHRandomlychooseA,BasthecentreandK=(1,)and(,)6cluster1cluster2newcenterABCDEFIGJHα(1,)β(,)<<>>>>><<<,βasthecentreandK=,weclassifyA,B,C,D,EasaclusterandF,G,H,I,(1,)β(,),βasthecentreandK=(,)andQ(,)8cluster2cluster1newcenterABCDEFIGJHP(,)Q(,)<<>>>>><<<,weclassifyA,B,C,D,EasaclusterandF,G,H,I,(,)Q(,)(,)andQ(,)P,QasthecentreandK=
kmeans算法的简单示例 来自淘豆网m.daumloan.com转载请标明出处.
