一种面向图书馆新书推荐服务的广义关联规则挖掘算法  

A Generalized Association Rule Mining Algorithm for Library New Book Recommendation

作　　者：佘俊胜[1] 黄战[1] 

机构地区：[1]暨南大学计算机科学系,广州510632

出　　处：《现代图书情报技术》2006年第10期16-20,共5页New Technology of Library and Information Service

摘　　要：基于MMS_Cumu late和GP-Apriori算法,提出一种针对图书馆新书推荐服务特点的广义关联规则挖掘算法MAR_LCR。不仅能挖掘出形如“读者-图书”的广义关联规则,而且还允许用户为不同的项设置不同的最小支持度。通过对候选集的产生过程进行改进,可大大压缩搜索空间。实验结果表明,MAR_LCR算法是有效的。最后,提出新书推荐模型。Based on MMS_Cumulate algorithm and GP - Apriori algorithm, a data mining algorithm, MAR_ LCR is proposed for library new book recommendation service which is capable of finding generalized association rules in the form of ＂reader- book＂ and allows the user to specify multiple minimum supports for different items. The search space is greatly cut down by improving the process of candidate generation. Experiment results show that the MAR_LCR algorithm is highly effective. Finally, a new book recommendation model is proposed.

关 键 词：广义关联规则 多最小支持度 新书推荐 图书馆 

分 类 号：G250.7[文化科学—图书馆学] TP311.13[自动化与计算机技术—计算机软件与理论]