An Autonomous Courses Recommender System For Undergraduate Using Machine Learning Techniques

Abstract

A University admission process is a complex process that needs a tremendous amount of time and labor to allocate course(s) to the prospective applicants. The study aimed at tackling the wrong placement of applicants into courses and to also address the wastage of admission vacancies by recommending the appropriate course(s). About 8,700 data for three academic sessions were collected from two different universities for training and testing the system. The features used are the results of Senior Secondary School subjects and the average score of UTME and Post-UTME. The proposed system employed five (5) classification models, which include Linear Regression, Naive Bayes, Support Vector Machine, K-Nearest Neighbor and Decision Tree Algorithm. The Linear Regression Model has the Root Mean Square Error (RMSE) as 2. 614 $e^{-14}$ The other four (4) classification models are also found to be efficient with an efficiency of at least 90% sequel to the dimensionality reduction in the dataset. The result shows that both Naive Bayes Classifier and Support Vector Machine have achieved the highest recommendation accuracy of approximately 99.94%, which outperformed Decision Tree and K-Nearest Neighbor algorithms with an accuracy of 98.10% and 99.87% respectively. The system can be adjusted with the change of admission criteria in Nigerian Universities. In consideration of the high degree of prediction accuracy, flexibility is an advantage, as the system can predict suitable courses that match the students’ grades. Therefore, the system is adaptive and good in making the prediction and can be used as a framework for further research on the admission recommender system.