METODE RCE-KMEANS UNTUK CLUSTERING DATA RCE-KMEANS METHOD FOR DATA CLUSTERING
| Main Authors: | , IZMY ALWIAH MUSDAR, , Dr. Azhari SN, M.T. |
|---|---|
| Format: | Thesis NonPeerReviewed |
| Terbitan: |
[Yogyakarta] : Universitas Gadjah Mada
, 2014
|
| Subjects: | |
| Online Access: |
https://repository.ugm.ac.id/134432/ http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=76978 |
Daftar Isi:
- Clustering is a data analysis technique that has been widely used in various fields. There have been many methods developed for solving clustering problems such as the methods in swarm intelligence field. Rapid Centroid Estimation (RCE) is one of Particle Swarm Optimization based method. RCE, like other variants of PSO clustering, does not depend on the initial centroid. RCE also has faster computitional time than the previous method like Particle Swarm Clustering (PSC) and the modified Particle Swarm Clustering (mPSC). However, its standard deviation of clustering scheme quality is higher than the PSC and mPSC which is influenced by variations of clustering results. This occurs because of improper equilibrium state, a condition in which the position of the particle does not change anymore, when stopping criteria is reached. RCE-Kmeans using K-means is used to solve equilibrium state problem in RCE method. K-means is applied after equilibrium state of RCE reached to update particle position result of RCE. There are 10 dataset used in this study. The quality of clustering scheme is measured by overall entropy, overall purity, and percentage misclassification. The quality of clustering scheme of RCE-Kmeans metode is compared with The quality of clustering scheme of RCE and K-means. The result of this study shows that RCE-Kmeans has better quality of the clustering scheme in 7 of 10 dataset compared to K-means and better in 8 of 10 dataset then RCE method. Moreover, using K-means in RCE can decrease standard deviation value of RCE method. Keyword : Clustering, Particle Swarm, K-means, Rapid Centroid Estimation