Analisis Cluster Non-Hirarki Dengan Menggunakan Metode K-Modes pada Mahasiswa Program Studi Statistika Angkatan 2015 FMIPA Universitas Mulawarman

Abstract

Cluster analysis is a technique that used to categorize or classify object into clusters or group which is relatively homogeneous. This research aims to know the number of the best cluster used in the selection of Statistics major using K-Modes Cluster, which variable as the best center of cluster & the most optimum, and also comparison of the cluster based on the Davies-Bouldin Index (DBI) which is derived in each cluster are 2 clusters, 3 clusters, and 4 clusters. Steps in this research is descriptive analysis, validity and reliability of questionnaire, determine the number of clusters, compute the dissimiliarity distance, calculate the cluster validation and interpretate the result of the best cluster. Selection of the best cluster use the smallest value comparison. The smallest of the two clusters are 0,599. The center (centroid) of clusters variables which is the best optimum using  K-Modes with two clusters are for the first centroid is the first choice of major, SNMPTN, IPK satisfactory, study routines for 4 times a week, and the average length of study is between 60 minutes to 120 minutes per day.; for the second centroid is the first choice of study program, SNMPTN, IPK is very satisfied, study routines for 6 times a week, and the average length of study is less than or equal to 60 minutes per day. The final results showed that the best cluster produced is two clusters where  cluster 1 consisted of 37 students and cluster 2 consisted of 8 students.