基于用户兴趣建模的个性化推荐  

PERSONALISED RECOMMENDATION BASED ON USERS' INTEREST MODELLING

作　　者：石林[1] 徐飞[1] 徐守坤[1] 

机构地区：[1]常州大学信息科学与工程学院,江苏常州213164

出　　处：《计算机应用与软件》2013年第12期211-214,264,共5页Computer Applications and Software

摘　　要：针对当前大多数个性化推荐中用户兴趣挖掘不足,导致资源推荐过快收敛的问题,以图书馆领域为背景,引入本体建模、本体查询、Apriori算法来全面挖掘用户潜在兴趣,同时利用概念频繁兴趣簇来控制最终用户推荐的收敛性。实验表明,该用户建模能保证在较高的资源推荐查准率基础上,防止推荐过快收敛,体现用户确切的兴趣。In light of the problems that in most personalised recommendation the users interest mining is insufficient currently, which leads to too fast convergence of the resources recommendation, in this paper, we take the field of library as the background, introduce the ontology modelling, ontology query and Apriori algorithm to comprehensively mine the potential users interest. Meanwhile, we use the concept frequent interest clusters to control the convergence of final users recommendation. Experimental results show that the user modelling can guarantee the prevention of too fast recommendation convergence on the basis of quite high resource recommendation precision rate, and reflect the exact interest of users.

关 键 词：本体查询 概念频繁兴趣簇 APRIORI算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]