BKG

Indriyani, Afrilia (2018) Penentuan Banyaknya Klaster Optimal Metode Average Linkage Menggunakan Gap Statistic, Silhouette, Davies-Bouldin. Sarjana thesis, Universitas Brawijaya.

Indonesian Abstract

Analisis klaster adalah teknik pengklasteran multivariat yang digunakan untuk mengklasifikasi objek atau individu ke dalam klaster yang sama berdasarkan karakteristiknya. Secara garis besar algoritma klaster dalam mengklasifikasi terdapat dua kategori, yaitu metode hirarki dan nonhirarki. Masalah utama dalam analisis klaster adalah menentukan banyaknya klaster optimal. Beberapa penelitian telah menghasilkan metode penentuan banyaknya klaster optimal menggunakan indeks validitas. Pada penelitian ini bertujuan untuk membandingkan tiga indeks validitas yaitu gap statistic, silhouette, dan Davies-Bouldin berdasarkan nilai Cluster Tightness Measure (CTM) terkecil dan melihat stabilitas klaster yang terbentuk menggunakan koefisien Jaccard. Terdapat empat data yang digunakan dalam penelitian ini. Berdasarkan hasil penelitian diperoleh kesimpulan bahwa indeks gap statistic merupakan indeks validitas terbaik karena menghasilkan nilai CTM terkecil.

English Abstract

Cluster analysis is a multivariate clustering technique used to classify objects or individuals into the same cluster based on their characteristics. Broadly clustered cluster algorithm there are two categories, namely hierarchy and non hierarchy method. The main problem in cluster analysis is to determine the optimal number of clusters. Several studies have resulted in the optimal cluster determination method using the validity index. This study aims to compare three validity indexes: statistic gap, silhouette, and DaviesBouldin based on the smallest Cluster Tightness Measure (CTM) value and see the cluster stability formed using Jaccard coefficients. There are four data used in this research. Based on the research results obtained conclusion that the index gap statistic is the best validity index because it produces the smallest CTM value.

Other Language Abstract

-

Item Type: Thesis (Sarjana)
Identification Number: SKR/MIPA/2018/340/051807393
Uncontrolled Keywords: Analisis Klaster, Cluster Tightness Measure (CTM), Indeks Validitas, Koefisien Jaccard, Cluster Analysis, Cluster Tightness Measure (CTM), Validity index, Jaccard Coefficients.
Subjects: 500 Natural sciences and mathematics > 519 Probabilities and applied mathematics > 519.5 Statistical mathematics > 519.53 Descriptive statistics, multivariate analysis, analysis of variance and covariance
Divisions: Fakultas Matematika dan Ilmu Pengetahuan Alam > Statistika
Depositing User: Nur Cholis
URI: http://repository.ub.ac.id/id/eprint/168658
Text
Afrilia Indriyani (2).pdf
Restricted to Repository staff only

Download

Actions (login required)

View Item View Item