PENERAPAN ANALISIS KLASTER K-MODES DENGAN VALIDASI DAVIES BOULDIN INDEX DALAM MENENTUKAN KARAKTERISTIK KANAL YOUTUBE DI INDONESIA (Studi Kasus: 250 Kanal YouTube Indonesia Teratas Menurut Socialblade)

Ahmad Badruttamam (Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro)
Sudarno Sudarno (Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro)
Di Asih I Maruddani (Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro)

Jurnal Gaussian
Vol 9, No 3 (2020): Jurnal Gaussian

YouTube is one of the most popular online platforms today. The popularity of YouTube has makes it an effective advertising medium. In April 2019, Socialblade released the top 250 YouTube channels in Indonesia based on their gradations with various characteristics. YouTube channel data will be grouped into several clusters to make it easier for advertisers to choose channels with characteristics as needed. The purpose of this study is to determine the best number of clusters and determine their characteristics. The method used is the k-Modes cluster analysis with values k = 3, 4, 5, ..., 8. The k-Modes method can group objects that have categorical type variables into relatively homogeneous groups. The best number of clusters (k) can be checked using the Davies Bouldin Index (DBI). Based on the analysis carried out, obtained the best number of six clusters with a Davies-Bouldin Index value of 1.080509. The most recommended cluster for advertising is cluster 6, which has grade A characteristics, gold title, and has an estimated annual income of 5 million USD < income ≤ 10 million USD. Keywords: Youtube, Cluster Analysis, k-Modes, Categorical Data, Davies-Bouldin Index