Comparison of DBSCAN and K-Means Clustering for Grouping the Village Status in Central Java 2020
| Main Authors: | Dewi, Cesaria, Siam, Emban Permata , Wijayanti, Gona Asri , Putri, Mustika , Aulia, Nurfitri , Nooraeni, Rani |
|---|---|
| Format: | Article info application/pdf eJournal |
| Bahasa: | eng |
| Terbitan: |
Department of Mathematics, Hasanuddin University
, 2021
|
| Subjects: | |
| Online Access: |
https://journal.unhas.ac.id/index.php/jmsk/article/view/11704 https://journal.unhas.ac.id/index.php/jmsk/article/view/11704/6711 |
Daftar Isi:
- Abstract Since Covid-19 was declared as a pandemic disaster, the world economic order has begun to be shaken, and Indonesia is no exception. Indonesia's economic growth has continued to contract since quarter II. Central Java Province is in the third place with the highest number of positive cases in Indonesia. The government try to improve the quality control over the implementation of village funds by observing the classification of village status. The status has been made by the Ministry of Villages based on IDM value. The purpose of this study is to create a village status cluster based on the three index values that compose the IDM, namely IKS, IKL, and IKE. This goal is realized through a comparative analysis of two clustering methods, that is K-means and DBSCAN. The results showed that by using the DBSCAN 4 clusters were formed, while using the K-means 3 clusters were formed. The silhouette value for each cluster formed using the DBSCAN is higher than the silhouette of clusters formed by the K-means and it is concluded that the DBSCAN is more appropriate to use in clustering village status in Central Java province in 2020 than K-means.
- Abstrak Sejak Covid-19 dinyatakan sebagai bencana pandemi, tatanan ekonomi dunia mulai terguncang, tidak terkecuali Indonesia. Pertumbuhan ekonomi Indonesia terus berkontraksi sejak triwulan II. Provinsi Jawa Tengah menempati urutan ketiga dengan jumlah kasus positif tertinggi di Indonesia. Pemerintah berupaya meningkatkan pengendalian mutu atas pelaksanaan dana desa dengan memperhatikan klasifikasi status desa. Status desa tersebut telah dibuat oleh Kementerian Desa berdasarkan nilai IDM. Penelitian ini bertujuan untuk membuat klaster status desa berdasarkan tiga nilai indeks penyusun IDM, yaitu IKS, IKL, dan IKE. Tujuan tersebut diwujudkan melalui analisis komparatif dari dua metode clustering, yaitu K-means dan DBSCAN. Hasil penelitian menunjukkan bahwa dengan menggunakan DBSCAN terbentuk 4 cluster, sedangkan dengan K-means terbentuk 3 cluster. Nilai silhouette untuk setiap cluster yang dibentuk menggunakan DBSCAN lebih tinggi daripada silhouette kluster yang dibentuk oleh K-means dan dapat disimpulkan bahwa DBSCAN lebih tepat digunakan pada clustering status desa di Provinsi Jawa Tengah pada tahun 2020 dibandingkan K-means.