Using Data Mining Techniques to Predict Student Performance to Aid Decision Making In University Admission Systems

Abstract

It is crucial that universities have a mechanism in place to choose students who will do well
academically based on objective factors used in the admissions process. This research looks
into how data mining tools can aid colleges in making admissions decisions by allowing for
more accurate predictions of student performance once they enrol. The proposed
methodology was tested on data from 2016-2019 involving 2,039 students at the Computer
Science and Information College of a Saudi state university. The findings show that
prospective students' early success in higher education can be forecast using data collected
before they are admitted using a variety of parameters (high school grade average, Scholastic
Achievement Admission Test score, and General Aptitude Test score). The results also
suggest that a student's score on the Scholastic Achievement Admission Test is the most
reliable indicator of future success. For this reason, this score needs to be given additional
weight in selection procedures. We also discovered that the Artificial Neural Network
method had a higher accuracy rate (over 79%) than the other categorization methods we
looked at (Decision Trees, Support Vector Machines, and Naive Bayes).