Comparison of Feature Selection Techniques for Predicting Student’s Academic Performance
- August 27, 2020
- Posted by: RSIS
- Category: IJRSI
International Journal of Research and Scientific Innovation (IJRSI) | Volume VII, Issue VIII, August 2020 | ISSN 2321–2705
Comparison of Feature Selection Techniques for Predicting Student’s Academic Performance
Olukoya, Bamidele Musiliu
Ph.D Student, Federal University Oye-Ekiti, Nigeria (FUOYE)
Abstract: In recent time, educational data mining (EDM) has received substantial considerations. Many techniques of data mining have been proposed to dig out out-of-sight knowledge in educational data. The Knowledge obtained assists the academic institutions to further enhance their process of learning and methods of passing knowledge to students. Powerful tools are required to analyze and predict the performance of students scientifically. This paper focuses on comparing two feature selection techniques in identifying major factors among the numerous affecting students’ academic that could give accurate prediction. Student educational data was retrieved from Kaggle data repository and feature selection on is done by applying Information Gain Attribute Evaluator and Correlation Based Features Selection (CFS) using WEKA as an Open Source Tool. Further a comparison is made among these two feature selections algorithm to select best attributes for prediction among all.
I. INTRODUCTION
The progress of a country is attached to the quality of its education system. There have been tremendous changes in educational setting across the globe in its functioning (Mishra, 2014). Like any other sector, education sector is facing challenges. The major challenges faced by higher education is abysmal students’ academic performance. To make the matter worse, some students leave school without completing their programs. One of the major objectives of any educational institution is to provide quality education to concerned students. Today, educational institutes and organizations seek to improve their systems by developing robust Information and Communication Technology (ICT) solutions to help the concerned managements in decisions making process. When this is done, it goes a long way to add value to the organizations objectively (Phil & Shoba, 2017).
In last decade, the number of higher education universities/institutions have proliferated manifolds. Large numbers of graduates/postgraduates are produced by them every year. Universities/Institutes may follow best of the pedagogies; but still they face the problem of dropout students, low achievers and unemployed students. Understanding and analyzing the factors for poor performance is a complex and incessant process hidden in past and present information congregated from academic performance and students’ behavior. Powerful tools are required to analyze and predict the performance of students scientifically. Although, Universities, Colleges/institutions collect enormous amount of students’ data before and after admitted into the school, but most of these data are unutilized.