Internal Cluster Validation on Earthquake Data in the Province of Bengkulu

K-means method is an algorithm for cluster n object based on attribute to k partition, where k < n. There is a deficiency of algorithms that is before the algorithm is executed, k points are initialized randomly so that the resulting data clustering can be different. If the random value for initi...

Full description

Saved in:
Bibliographic Details
Published inIOP conference series. Materials Science and Engineering Vol. 335; no. 1; pp. 12048 - 12056
Main Authors Rini, D S, Novianti, P, Fransiska, H
Format Journal Article
LanguageEnglish
Published Bristol IOP Publishing 01.04.2018
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:K-means method is an algorithm for cluster n object based on attribute to k partition, where k < n. There is a deficiency of algorithms that is before the algorithm is executed, k points are initialized randomly so that the resulting data clustering can be different. If the random value for initialization is not good, the clustering becomes less optimum. Cluster validation is a technique to determine the optimum cluster without knowing prior information from data. There are two types of cluster validation, which are internal cluster validation and external cluster validation. This study aims to examine and apply some internal cluster validation, including the Calinski-Harabasz (CH) Index, Sillhouette (S) Index, Davies-Bouldin (DB) Index, Dunn Index (D), and S-Dbw Index on earthquake data in the Bengkulu Province. The calculation result of optimum cluster based on internal cluster validation is CH index, S index, and S-Dbw index yield k = 2, DB Index with k = 6 and Index D with k = 15. Optimum cluster (k = 6) based on DB Index gives good results for clustering earthquake in the Bengkulu Province.
ISSN:1757-8981
1757-899X
DOI:10.1088/1757-899X/335/1/012048