Course Selection Optimization: Case study - Faculty of Science, University of Peradeniya, Sri Lanka

Abstract

Elective course selection at universities is a complex decision process which is subjective to each individual student’s personality and skill set. The aim of this research is to use machine learning techniques and expert knowledge to suggest optimal course selections by considering the student skills (profile of the student) and the profiles of the courses offered at the university. It takes into consideration the fact that if a student is performing well in a particular course, he/she can select another course of the same nature to improve the student’s results and give a solution to the daunting task of selecting elective courses. The K-Nearest Neighbour algorithm resulted in ten course clusters for the dataset and accordingly students were grouped using the highest average course cluster GPA. Application of the expert knowledge method resulted in course clusters which can be split into clusters as stipulated by the Faculty. The approach was validated for computer science courses offered at the Faculty of Science, University of Peradeniya, Sri Lanka, as a case study from 2005 to 2012.