K-Medoid Algorithm in Clustering Student Scholarship Applicants

Abstract

Data Grouping scholarship applicants Bantuan Belajar Mahasiswa (BBM) grouped into 3 categories entitled of students who are eligible to receive, be considered, and not eligible to receive scholarship. Grouping into 3 groups is useful to make it easier to determine the scholarship recipients fuel. K-Medoids algorithm is an algorithm of clustering techniques based partitions. This technique can group data is student scholarship applicants. The purpose of this study was to measure the performance of the algorithm, this measurement in view of the results of the cluster by calculating the value of purity (purity measure) of each cluster is generated. The data used in this research is data of students who apply for scholarships as many as 36 students. Data will be converted into three datasets with different formats, namely the partial codification attribute data, attributes and attribute the overall codification of the original data. Value purity on the whole dataset of data codification greatest value is 91.67%, it can be concluded that the K-Medoids algorithm is more suitable for use in a dataset with attributes encoded format overall.