Journal of International Conference Proceedings
Vol 6, No 5 (2023): 2023 UICEB Papua Proceeding

Performance of K-Means and DBSCAN Algorithm in Clustering Gross Regional Domestic Product

Wororomi, Jonathan K. (Unknown)
Allo, Caecilia Bintang Girik (Unknown)
Paranoan, Nicea Roona (Unknown)
Gusthvi, Wickly (Unknown)



Article Info

Publish Date
13 Dec 2023

Abstract

Gross Regional Domestic Product (GRDP) is one of important indicator to determine the economic conditions of a region. GRDP are obtained from sum of value added produced by all unit of production in a region. This study use GRDP by production approach that grouped into seventeen categories of Industry. The government always put the big efforts to increase the economic growth after Covid-19 pandemic. According publication of BPS - Statistics Indonesian, in the year of 2021 and 2022 it’s growth between 3.70% and 5.31%. The aim of these study are determined the cluster GDRB based on province in Indonesia at current prices and analyses the performance of the cluster method. The results showed that by using the DBSCAN, two clusters were formed and one province can be detected as an outlier. On the other hands, performance of the method by K-Means showed two clusters. The silhouette value using K-Means is higher than the DBSCAN. For these case, the performance of K-Means is more appropriate than DBSCAN to use in clustering province in Indonesia based on GRDP at Current Market Prices. Moreover, performance of DBSCAN shows more sensitive on outliers detection.

Copyrights © 2023






Journal Info

Abbrev

JICP

Publisher

Subject

Humanities Decision Sciences, Operations Research & Management Electrical & Electronics Engineering Industrial & Manufacturing Engineering Languange, Linguistic, Communication & Media

Description

JICP is proceedings series that aims to publish proceedings from conferences, in the fields of economics, business, and management research. All proceedings in this website are open access, which means the published articles are permanently free to read, download, copy, and distribute. The online ...