Analisis Cluster Pada Data Kategorik dan Numerik dengan Pendekatan Cluster Ensemble

Studi Kasus: Puskesmas di Provinsi Kalimantan Timur Kondisi Desember 2017

  • Nur Aini Ayu Lestari Laboratorium Statistika Komputasi FMIPA Universitas Mulawarman
  • Memi Nor Hayati Laboratorium Statistika Terapan FMIPA Universitas Mulawarman
  • Fidia Deny Tisna Amijaya Laboratorium Matematika Komputasi FMIPA Universitas Mulawarman

Abstract

Cluster analysis used to process categorical and numerical data at once is Cluster Ensemble algorithm Based on Mixed Data Clustering (algCEBMDC), which is a cluster algorithm with an ensemble cluster approach. The method used for numerical data is Agglomerative Nesting (AGNES) algorithm and for categorical data is the RObust Clustering using linK (ROCK) algorithm. The best clustering method and the optimum number of clusters in the AGNES algorithm is selected based on the maximum Pseudo-F value and the minimum icdrate value. The optimum number of clusters in the ROCK algorithm is selected using the minimum value of  ratio . The purpose of this study was to make a group of 179 Puskesmas in East Kalimantan on December 2017. Based on the results of the analysis, obtained 5 optimum cluster for numerical clustering with the AGNES algorithm and 2 optimum cluster for categorical clustering data with the ROCK algorithm. Final cluster for mixed data clustering obtained 2 optimum cluster at a threshold of 0.2 and 0.3 with value of ratio  is . The first cluster consists of 83 Puskesmas and cluster two of 96 Puskesmas.

Downloads

Download data is not yet available.
Published
2021-01-19
How to Cite
LESTARI, Nur Aini Ayu; HAYATI, Memi Nor; AMIJAYA, Fidia Deny Tisna. Analisis Cluster Pada Data Kategorik dan Numerik dengan Pendekatan Cluster Ensemble. EKSPONENSIAL, [S.l.], v. 11, n. 2, p. 117-126, jan. 2021. ISSN 2798-3455. Available at: <https://jurnal.fmipa.unmul.ac.id/index.php/exponensial/article/view/652>. Date accessed: 29 mar. 2024.
Section
Articles