A preference-based recommendation method with fuzzy comprehensive evaluation

This paper proposes a preference-based recommendation method with fuzzy comprehensive evaluation to support personalization. Unlike the conventional approach in which only the users' purchasing behavior and browsing records can be used, the proposed recommendation method analyzes users' current preference captured from the users' input, and calculates the weights of product attributes and the fuzzy comprehensive evaluation with users' evaluation matrix. Rank the products in descending order according to evaluation value for recommendation. Weight calculation and fuzzy comprehensive evaluation are the two key technologies, with which the method can be easily carried out. The feasibility of the proposed method is validated with a case. Also more experiments are tested to validate our method. You can also see comparisons between manual and computer recommendations. The experimental results show that applying fuzzy comprehensive evaluation is effective and accurate.