K-Means Clustering Algorithm in Web-Based Applications for Grouping Data on Scholarship Selection Results

Abstract

In our case the selection of scholarship recipients was done previously by a university, by assessing and scoring 7 variables: whether scholarship application form is submitted, whether student study plan is submitted, whether student transcript is submitted, whether student's curriculum vitae is submitted, what is the score of student's scientific article, what is the score of student's presentation on a given topic, and what is the score of student's GPA. This such selection process was done without computation, with the consequence of time consumption and potential human errors. This research aims to overcome that problem, by providing a computed selection process using the same 7 variables and applying a necessary algorithm. In our research the K-means Clustering algorithm is applied although it is understood that other algorithms can be used too. The tests are carried out using black box and white box methods. The result shows that K-means Clustering algorithm is successfully applied to the scholarship selection system, and the K-means Clustering algorithm is successful in grouping students who receive and who do not receive the scholarship.