Classification of bandung regency well water based on mineral elements using agglomerative clustering

¹ Statistics Research Division, Faculty of Mathematics and Natural Science, Bandung Institute of Technology, 10 th Ganesha Street, Bandung, Indonesia, 40132
² Doctoral Program in Mathematic, Faculty of Mathematics and Natural Science, Bandung Institute of Technology, 10 th Ganesha Street, Bandung, Indonesia, 40132
³ Undergraduate Program in Mathematics, Faculty of Mathematics and Natural Science, Bandung Institute of Technology, 10 th Ganesha Street, Bandung, Indonesia, 40132
⁴ Graduate Program in Computational Science, Faculty of Mathematics and Natural Science, Bandung Institute of Technology, 10 th Ganesha Street, Bandung, Indonesia, 40132
⁵ Environmental Management Technology Research Group, Department of Environmental Engineering, Faculty of Civil and Environmental Engineering, Bandung Institute of Technology, 10 th Ganesha Street, Bandung, Indonesia, 40132

^* Corresponding author: yuli.afrianti@itb.ac.id

Abstract

Wells are crucial sources of drinking water, but their safety depends on the mineral element, making quality standards assessments. This study evaluates well water quality across 160 locations in Bandung Regency by analyzing six mineral elements (i.e., As, Cd, Fe, Mn, Pb, and Zn) using Agglomerative Clustering with different linkages (i.e., Single, Average, and Complete) and distance metrics (i.e., Euclid and Manhattan). The aim of this study is to review the distribution of well quality using the Agglomerative clustering method which represents the Bandung Regency region. The optimal number of clusters is determined via the Mojena method, and the best linkage is selected using the Silhouette Coefficient. The study finds that Average Linkage with Euclid distance metrics and Single Linkage with Manhattan distance metrics are the most effective methods. We then assess each cluster against drinking water standards to determine quality levels, which range from 1 (poor) to 4 (excellent). The results indicate that Average Linkage with Euclid distance metrics better represents well water data. These findings are crucial for guiding the management of safe water resources based on regional characteristics.