Quality enhancement at higher education institutions by early identifying students at risk using data mining

Abstract

Accurate prediction of students' academic performance is one of the challenges in maintaining quality standards in any Higher Education Institution (H.E.I.). To ensure the quality of teaching and learning, H.E.I.s often employ Self-Assessment Reports (S.A.R.s) in which identifying a student drop-out ratio is important. Hence, it is essential to identify at-risk students in a given academic program. This article aims to identify at-risk students early by proposing a data mining-based predictive framework to improve the student's learning experience and minimize the dropped-out ratio. The academic sub-attributes or indicators in each course that may affect the performance of students in higher education institutions used in this study to examine students' academic achievement and predict students' performance to distinguish at-risk students are the marks of assignments, mid-term, lab exams, semester marks, total, grade, grade point (G.P.), quality point (Q.P.), grade point average (G.P.A.), and credit hours data of multiple courses categorized according to three knowledge areas defined by Higher Education Commission (H.E.C), Pakistan using data mining predictive techniques. The results indicate that the proposed methods can achieve maximum accuracy in predicting and identifying at-risk students in different courses.

Dimensionality Reduction using Feature Selection Techniques on EDM for Student Academic Performance Prediction

“…For this, the results were evaluated using various statistical techniques (STATS), including frequency-counts, percentages, averages, weighted averages, and T-tests [20]. In order to enhance students' learning experiences and reduce the dropped-out ratio [21], the authors tried to find at-risk students early on by suggesting a prediction framework based on data mining. Various data mining techniques such as K-Neural Network, Decision Tree with gain ratio, Naïve Bayes, Random Forest with gain ratio with logistic regression were employed.…”

Comparing Different Oversampling Methods in Predicting Multi-Class Educational Datasets Using Machine Learning Techniques

“…Additionally, educational data mining can provide valuable insights into the effectiveness of online and blended learning environments, helping to guide the development of these educational formats. Overall, educational data mining has the potential to revolutionize the way we approach education and has the potential to have a significant impact on student outcomes and the future of education [2][3][4].…”