Abstract: According to the weak association rule problems which caused by sparse association rule, and recommendation problems such as the lack of diversity, the dynamic product taxonomy is generated based on Vague set theory, and mining association rule in product taxonomy is presented to solve the weak association rules problem; on this basis, a diversity selection algorithm is proposed to solve recommendation diversity problem in recommendation result set. Experimental evaluation results show that both in the recommendation accuracy and recommendation diversity is more effective compared with the traditional recommendation method.

Key words: recommendation system, association rules, product taxonomy, Vague sets theory, recommendation diversity

摘要： 针对关联规则过于稀疏导致的弱关联规则问题，以及关联规则推荐存在的多样性匮乏等问题，提出基于Vague理论生成动态产品分类树，在分类树内实施关联规则挖掘以解决弱关联规则问题；在此基础上进一步提出一种基于产品相似性的多样性选择算法，并在推荐结果集内实施多样性选择以解决推荐多样性问题，实验评价结果表明该方法与传统推荐方法相比，无论在推荐精度还是推荐多样性上都更为有效。

关键词: 推荐系统, 关联规则, 产品分类树, Vague集理论, 推荐多样性