<![CDATA[ne"> Abstract— The disadvantage of the K-means algorithm is sensitiveto have problems determining the initial partition number ofclusters (k) determining the initial value that is different mayproduce different cluster groups. To solve the problem of thesensitivity of the initial partition number of clusters in K-meansalgorithm, the proposed algorithm dynamic cluster. The resultshowed that the Dynamic K-means algorithm, can produce qualitycluster that is more optimal than the K-means.Intisari— Salah satu kekurangan algoritma K-means yaitumempunyai masalah sensitif terhadap penentuan partisi awaljumlah cluster(k) penentuan nilai awal yang berbeda mungkindapat menghasilkan kelompok cluster yang berbeda pula. Untukmenyelesaikan masalah sensitifitas partisi awal jumlah clusterpada algoritma K-means, maka diusulkan algoritma clusterdinamik. Hasil percobaan menunjukan bahwa algoritmaDynamic K-means, dapat menghasilkan kualitas cluster yanglebih optimal dibandingkan dengan K-means.Kata kunci : Segmentasi Pelanggan, K-Means, quality cluster]]>
Copyrights © 2015
