一个基于演化关联规则挖掘的个性化推荐模型  

A personalized recommendation model based on evolving association rule mining

作　　者：龙舜[1] 蔡跳[1] 林佳雄[1] 

机构地区：[1]暨南大学计算机科学系,广东广州510632

出　　处：《暨南大学学报（自然科学与医学版）》2012年第3期264-267,共4页Journal of Jinan University(Natural Science & Medicine Edition)

基　　金：武汉大学软件工程国家重点实验室开放基金项目(SKLSE 2010-08-31);国家科技计划重大专项课题(2008ZX10005-013);广东省科技计划项目项目(2010A032000002)

摘　　要：关联分析是一种非常有效的数据挖掘方法,它能帮助人们发现隐藏在大型数据集中的令人感兴趣的联系,在电子商务等应用领域取得了广泛的应用.但实际运用中,它仍存在着关联规则挖掘困难以及得到的静态结果不易及时反映情况变化等问题.为此,人们提出演化关联规则的概念,本文设计了一个基于演化规则集的推荐模型.实际测试表明:它可挖掘出有用的关联规则并及时反映情况的变化,为客户提供更到位的个性化推荐服务,具有简单和计算效率高等特点.Association analysis is an effective data mining approach capable of unveiling interesting associations within a large dataset. Although widely adopted in e-business areas, it still has difficulties in practice. For instance, the resulting rules appear static and therefore cannot reflect in-time the varying nature of people＇ s interests. The concept of evolving association rule is introduced to tackle these problems. This paper presents our recommendation model based on evolving association rule mining. Experimental results on an online toggery show that it can effectively unveil people＇ s varying interests, and the mining process is simple and efficient.

关 键 词：数据挖掘 演化关联规则 个性化推荐 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]