Clustering the College Admission Test Performance Using K-Means, Hierarchical Clustering Class and Expectation Maximization Algorithms

Authors

  • Rosalinda B. Guiyab

Abstract

Clustering the College Admission Test (CAT) can be very helpful in assisting colleges and universities in their
admissions. The scores from these tests will help assess students for admission and help make admission decisions. This study
utilized 3 clustering algorithms namely K-Means, Hierarchical Clustering Class, and Expectation Maximization algorithms applied to
CAT data. It aimed to determine which clustering algorithm is best in determining the CAT performance and further using the best
clustering algorithm to establish groups with respect to the High School (HS) last attended by the student. This is to determine the
performance attributes of the group that can be used to improve performance in the CAT in the future and consequently, evaluate and
compare the preparation of students who go to different High Schools which can be helpful to improve decision making. Result of the
study showed that the best clustering algorithm applied to CAT performance data is the K-means algorithm.

Published

2020-04-30

Issue

Section

Articles