Implementation of the K-Means Algorithm for Clustering the Characteristics of Students Receiving Kartu Indonesia Pintar Kuliah (KIP-K)

Abstract

The limited quota of recipients of the Kartu Indonesia Pintar Kuliah (KIP-K) causes the host universities to select applicants to get students who are eligible to receive KIP-K based on academic achievement, non-academic achievements, and family economic conditions. However, after the lectures started, some students who received KIP-K lacked discipline in undergoing lecture procedures and experienced a decrease in their achievement index (IP). Therefore, it is necessary to explore knowledge about the characteristics of KIP-K recipient students by conducting clustering modeling. So, in this study, clustering modeling was carried out on student data receiving KIP-K at a university by applying the Cross-Industry Standard Process for Data Mining (CRIPS-DM) method and the k-means clustering algorithm. This study chooses a clustering model with a value of k=2, which has the smallest Davies Bouldine index (DBI) value of 0.35. This clustering resulted in 2 clusters where student characteristics showed significant differences in the attributes of the distance from home to the campus location and relatively minor fluctuations in IP from the first semester to the fourth semester. From mapping the characteristics of KIP-K recipient students, knowledge can be used as material for higher education decisions in selecting KIP-K registrants to minimize the future academic problems of KIP-K recipient students.