IT Students Selection and Admission Analysis using Naïve Bayes and C4.5 Algorithm

Abstract

© 2020, World Academy of Research in Science and Engineering. All rights reserved. Admission to college and selection of applications have probably become an integral part of some colleges and universities in their enrolment process, yet it is girded by controversy and skepticism. A new area of research that uses techniques of data mining is known as Educational Data Mining. It incorporates machine learning algorithms and statistical methods to help for the interpretation of student’s learning habits, academic performances, and further improvements-if needed. This paper focuses on the predictive values of certain academic variables, admission tests, high school academic records as related to the performance of Information Technology (IT) students at the end of the first year. For this reason, 221 data were used, and C4.5 and Naive Bayes algorithms are applied to generate a prediction on the students' performance. The C4.5 classification gained 98.64% in 10-folds cross-validation and 96.97% in the 70% training and 30% testing percentage split compared to Naïve Bayes which only gained 89.14% and 86.36% for both 10-folds cross-validation and 70% training and 30% testing percentage split respectively. The comparative analysis of the result shows that senior high school track and academic data and admission test results are the influential attributes to the performance of IT students in their first year. This paper recommends for future studies to add different data from different years to increase the accuracy of the prediction.