Please use this identifier to cite or link to this item:
|Title:||Using machine learning to support students’ academic decisions|
|Authors:||ALLAH, AISHA QASIM GHAZAL FATEH|
|Publisher:||The British University in Dubai (BUiD)|
|Abstract:||Making the right decision for students in higher education is vital, as it has a great influence on their study, career, life, and eventually, the whole society. Predicting the future performance of students can inform their choice of majors, concentrations, and courses. It also helps teachers and advisors provide the necessary support to students as needed. While many studies address the issue of predicting students’ performance, they mainly predict student performance at one stage of their study only. This work proposes a framework for assisting students in their decision throughout their study journey. At enrollment, this work predicts a student’s GPA in different majors using enrollment data such as high school average, placement test results, and IELTS score. After completing their first year, this work predicts student’s GPA in different concentrations using grades of Year 1 courses. At any point of time after the student finishes some courses, a user-based collaborative filtering approach using K-Nearest Neighbor is used to predict a student’s grade in a future course. This approach uses other students’ grades to make a prediction. This research tests and compares the performance of Decision Trees, Random Forests, Gradient-Boosted trees, and Deep Learning machine learning regression algorithms to predict student GPA. Furthermore, the strongest predictors of student’s GPA are identified at each stage. Gradient Boosted Trees performed the best when predicting student’s Major GPA, while Deep Learning performed the best for predicting Concentration’s GPA.|
|Appears in Collections:||Dissertations for Informatics (Knowledge and Data Management)|
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.