A Hybrid Programming Course recommendation system using Fuzzy Logic and xDeepFM

Abstract

The choice of courses is crucial to a student's career achievement. Different courses offered by various academic institutions need the students to browse the course outline in the current educational environment manually. Most of them need more understanding, struggle with decisions, and make rash decisions regarding the best course of action. A course recommendation system helps students choose a course that suits their interests and abilities. The skill level classification using Mamdani FIS allows a new student to select an equivalent course that can solve the cold start issue of the recommendation system. After the student chooses the initial course, the xDeepFM model can continue recommending relevant courses to help the student. This research contributes to integrating the FIS with xDeepFM to enhance the accuracy of course recommendations. Additionally, the hybrid method solves the cold start and the data sparsity problem and simultaneously lessens variability in relevant course search results and information overload. The experimentation result indicates that xDeepFM is a suitable method to be combined with FIS to build an efficient hybrid course recommendation system.