TINJAUAN PUSTAKA SISTEMATIS PERBANDINGAN ALGORITMA K-MEANS DAN DBSCAN PADA KLASTERISASI DATA LAYANAN PUBLIK

Systematic Literature Review of K-Means And DBscan Algorithms Comparison In Public Service Data Clustering

Abstract

This study presents a systematic literature review comparing K-Means and DBSCAN in clustering public service data. The article adopts the PRISMA approach to make the review process systematic, transparent, and reproducible. Searches were conducted in Google Scholar, ScienceDirect, and IEEE Xplore using keywords related to K-Means, DBSCAN, clustering, and public service data. A total of 11,700 records were identified, 140 studies were retained as candidates after initial screening, and 30 studies were included in the final synthesis. The review shows that K-Means is often used because it is simple, fast, and efficient for structured data, while DBSCAN performs better for datasets containing noise, density variation, and irregular cluster shapes. In public service data, algorithm selection depends on data characteristics, analytical objectives, and parameter sensitivity.