Hybrid Recommender Systems for Course Selection Using Collaborative Filtering and Multi-Criteria Decision-Making

The rapid growth of Massive Open Online Course (MOOC) catalogs has made it increasingly difficult for learners to identify courses that are simultaneously aligned with their personal interests and objectively of high quality. Single-technique recommender systems tend to address only one of these two dimensions: collaborative filtering (CF) captures personalized preference but ignores objective course attributes, whereas multi-criteria decision-making (MCDM) evaluates objective quality but disregards individual taste. This study proposes and demonstrates a hybrid recommender system that integrates User-Based Collaborative Filtering with Cosine Similarity and the Simple Additive Weighting (SAW) method to generate course recommendations that balance personalization with objective merit. A dummy dataset consisting of 10 MOOC items, 8 users, and 5 evaluation criteria (average rating, relevance to interest, difficulty level, duration, and price) was constructed to test the proposed method end-to-end. The Collaborative Filtering component predicted preference scores for a target user through similarity-weighted aggregation of the top-3 nearest neighbors, while the SAW component produced an objective preference value for each course through max-normalization and weighted summation across the five criteria. The two outputs were linearly combined using an equal-weighted hybridization coefficient (? = 0.5) to produce a final ranked list. Results show that the hybrid model reorders the recommendation list relative to either technique used alone, promoting courses that are both well-matched to the target user's demonstrated preferences and objectively strong on relevance, quality, and cost-efficiency. The proposed pipeline is technique-agnostic and can be re-parameterized for other selection domains that require a combination of personalized filtering and objective multi-criteria evaluation