Abstract
This research addresses the challenges of sparsity and uncertainty in user-product interaction data by integrating fuzzy logic with collaborative filtering. The proposed fuzzy CF framework utilizes cosine similarity metrics and triangular membership functions to refine similarity scores and predict ratings for unrated items. Experimental results demonstrate that while the fuzzy CF system slightly increases RMSE, it significantly enhances recommendation coverage, making it robust in sparse data scenarios. The findings suggest that fuzzy logic effectively complements traditional CF methods to improve recommendation quality and coverage.