Comparative Analysis of Sensitivity and Performance of K-Means and DBSCAN to Noise Variations

• Rafael Austin

  Universitas Esa Unggul

• Ricky Dwi Putra

  Esa Unggul University

• Ardi Ardi

  Esa Unggul University

• Fuzail Fazle Rabbi

  Esa Unggul University

• Lucky Pujiono WS

  Esa Unggul University

• Vitri Tundjungsari

  Esa Unggul University

Inaccurate or noisy data presents a significant challenge in machine learning, particularly in unsupervised clustering tasks. This study evaluates the robustness and performance of two popular clustering algorithms, K-Means and DBSCAN, against various levels of Gaussian noise (5%, 10%, 20%, and 30%) injected into a customer dataset. Evaluation was conducted using Silhouette Score and Davies-Bouldin Index (DBI). Initial results indicated that DBSCAN performed slightly better with a Silhouette Score of 0.4817 compared to K-Means at 0.4101. However, after noise injection, K-Means demonstrated superior stability by maintaining more consistent cluster memberships, whereas DBSCAN was more sensitive to distance variations, leading to significant fluctuations in cluster assignments. The study concludes that K-Means is more reliable for datasets where cluster integrity must be preserved despite minor data irregularities.

[1] J.-Y. Franceschi, A. Fawzi, and O. Fawzi, “Robustness of classifiers to uniform p and Gaussian noise,” 2018.

[2] S. Gupta and A. Gupta, “Dealing with noise problem in machine learning data-sets: A systematic review,” in Procedia Computer Science, Elsevier B.V., 2019, pp. 466–474. doi: 10.1016/j.procs.2019.11.146.

[3] S. Pau, A. Perniciano, B. Pes, and D. Rubattu, “An Evaluation of Feature Selection Robustness on Class Noisy Data,” Information (Switzerland), vol. 14, no. 8, Aug. 2023, doi: 10.3390/info14080438.

[4] Y. Sun and J. Zhang, “Distance-based Data Cleaning: A Survey (Technical Report),” Nov. 2020, [Online]. Available: http://arxiv.org/abs/2011.11176

[5] C.-H. Lin, E. L. Dyer, V. Muthukumar, and C. Kaushik, “The good, the bad and the ugly sides of data augmentation: An implicit spectral regularization perspective,” 2024.

[6] Tundjungsari Vitri, “Dasar Machine Learning”.

[7] P. Khairani Ritonga and M. S. Hasibuan, “Analisis Perbandingan Silhouette dengan Elbow pada Algoritma K-Means dan DBSCAN,” vol. 9, p. 2025, doi: 10.47002/metik.v9i1.1027.

[8] P. Vania and B. Nurina Sari, “Perbandingan Metode Elbow dan Silhouette untuk Penentuan Jumlah Klaster yang Optimal pada Clustering Produksi Padi menggunakan Algoritma K-Means,” Jurnal Ilmiah Wahana Pendidikan, vol. 9, no. 21, pp. 547–558, 2023, doi: 10.5281/zenodo.10081332.

[9] R. C. de Amorim and V. Makarenkov, “Improving clustering quality evaluation in noisy Gaussian mixtures,” Mar. 2025, [Online]. Available: http://arxiv.org/abs/2503.00379

[10] T. Setyani, Y. Indriani, M. Fadli, and E. R. Susanto, “Perbandingan Algoritma K-Means dan DBSCAN dalam Pengelompokkan Data Penjualan Elektronik,” Progresif: Jurnal Ilmiah Komputer, vol. 21, no. 2, p. 712, Aug. 2025, doi: 10.35889/progresif.v21i2.2775.

[11] C. H. Lin, K. C. Hsu, K. R. Johnson, M. Luby, and Y. C. Fann, “Applying density-based outlier identifications using multiple datasets for validation of stroke clinical outcomes,” Int. J. Med. Inform., vol. 132, Dec. 2019, doi: 10.1016/j.ijmedinf.2019.103988.

[12] F. Andriyani and Y. Puspitarani, “Performance Comparison of K-Means and DBScan Algorithms for Text Clustering Product Reviews,” SinkrOn, vol. 7, no. 3, pp. 944–949, Jul. 2022, doi: 10.33395/sinkron.v7i3.11569.

[13] R. Adesunkanmi and R. Kumar, “Expectation Distance-based Distributional Clustering for Noise-Robustness,” Mar. 2023, [Online]. Available: http://arxiv.org/abs/2110.08871

[14] O. Ajmal, H. Arshad, M. A. Arshed, S. Ahmed, and S. Mumtaz, “Robust Parameter Optimisation of Noise-Tolerant Clustering for DENCLUE Using Differential Evolution,” Mathematics, vol. 12, no. 21, Nov. 2024, doi: 10.3390/math12213367.

[15] Shabila Ocktavia and Wahyu Tisno Atmojo, “ANALISIS SEGMENTASI PELANGGAN PEMBIAYAAN BERDASARKAN DEMOGRAFI UNTUK MEMPREDIKSI TINGKAT KREDIT MENGGUNAKAN ALGORITMA K-MEANS”.

• PDF

Copyright (c) 2026 Rafael Austin, Ricky Dwi Putra, Ardi Ardi, Fuzail Fazle Rabbi, Lucky Pujiono WS, Vitri Tundjungsari

This work is licensed under a Creative Commons Attribution 4.0 International License.