Hidayati, Rahmatina, Zubair, Anis, Pratama, Aditya Hidayat and Indana, Luthfi ORCID: https://orcid.org/0000-0003-3270-0143 (2021) Analisis silhouette coefficient pada 6 perhitungan jarak K-Means Clustering. Tecno.com Jurnal Teknologi Informasi, 20 (2). pp. 186-197. ISSN E-ISSN : 2356-2579 P-ISSN : 1412-2693

Preview	Text Analisis Silhouette Coefficient pada 6 Perhitungan Jarak.pdf Download (531kB) | Preview
Preview	Text Hasil Cek Plagiasi.pdf Download (1MB) | Preview

Abstract

Clustering is the process of grouping a set of data into clusters that have similarities. The
similarity in a cluster is determined by calculating the distance. To see the performance of some
distance calculations, in this study the authors tested 6 data which had different attributes,
namely 2, 3, 4, and 6 attributes. From the results of the comparison test for the distance formula
on K-Means clustering using the Silhouette coefficient, it can be concluded that: 1) Chebyshev
distance has stable performance both for data with few and many attributes. 2) Average
distance has the highest Silhouette coefficient result compared to other distance measurements
for data that has outliers such as data 3. 3) Mean Character Difference gets good results only
for data with few attributes. 4) Euclidean distance, Manhattan distance, and Minkowski
distance produce good values for data that have few attributes, while data with many attributes
get a sufficient value that is close to 0.5.

Actions (login required)

View Item