用户聚类和多最小支持度关联规则的推荐系统

在几种流行的推荐算法的基础上，本文提出了一种新的融合用户聚类和关联规则的算法来改善推荐效果。该算法在经典的关联规则算法Apriori基础上引入多最小支持度的概念，并在关联规则算法之前进行用户聚类，在聚类算法中使用了包含字符属性的混合属性距离函数，提高聚类效果。在此算法的基础上，设计并实现了一种新的基于图书馆的推荐系统。实验证明该算法改善了数字图书馆中新书的推荐质量，去除了部分只含高浏览量图书的无意义规则，并趋向于发现相近种类图书的关联性。

A recommendation system combining user clustering and association rules with multiple minimum support

On the basis of several popular recommendation algorithms, a new algorithm combining user clustering and association rules is proposed to improve the efficacy of a recommendation system. This algorithm is based on the classic association rules and the Apriori algorithm, and introduces the concept of minimum support. Then, prior to the association rules, it carries out user clustering. The mixed attribute distance function including character attributes used in the clustering algorithm can improve the clustering precision. Finally, a new recommendation system is designed and implemented for library users. The experimental results show that the proposed algorithm improves the quality of recommendations of new library books, reduces the quantity of insignificant rules which only include widely browsed books, and tends to flag the relevance of books of similar types.