The Effect of Preliminary Centroid Determination Using Particle Swarm Optimization Algorithm in High Dimension Data Clustering

DOI

10.2991/aisr.k.200424.052How to use a DOI?

Keywords

particle swarm optimization, principal component analysis, clustering, k-means, davies bouldin index

Abstract

K-Means is one of the methods in clustering where the results are strongly influenced by initial centroid positioning. In general the k-means method in determining the initial centroid is generated randomly. Determination of the initial random centroid is what often makes k-means trapped in the optimum local solution that results in poor cluster quality. This research will examine the effect of the particle swarm optimization algorithm in determining the initial centroid of the k-means method. Based on the results of the k-means clustering test with the initial centroid of particle swarm optimization can improve the quality of the cluster, whether tested on data reduction or without reduction, with a percentage change of 43.8% in data without dimension reduction and 53.4% in dimensionally reduced data.

Copyright

© 2020, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Ardi Wasila CHANDRA
AU  - Rifkie PRIMARTHA
AU  - Anita Desiani
AU  - Adi WIJAYA
PY  - 2020
DA  - 2020/05/06
TI  - The Effect of Preliminary Centroid Determination Using Particle Swarm Optimization Algorithm in High Dimension Data Clustering
BT  - Proceedings of the Sriwijaya International Conference on Information Technology and Its Applications (SICONIAN 2019)
PB  - Atlantis Press
SP  - 347
EP  - 352
SN  - 1951-6851
UR  - https://doi.org/10.2991/aisr.k.200424.052
DO  - 10.2991/aisr.k.200424.052
ID  - CHANDRA2020
ER  -