基于AOI的客户行为分析方法  

CUSTOMER BEHAVIOR ANALYSIS ALGORITHM BASED ON ATTRIBUTE-ORIENTED INDUCTION

作　　者：薛军[1] 陈英[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《计算机应用与软件》2008年第6期126-127,152,共3页Computer Applications and Software

摘　　要：结合数据立方体技术以及概念分层的分析方法,将面向属性的归纳方法(AOI)与K-means聚类算法相结合,应用于客户时序数据聚类分析中,使每一类客户都具有相似的时序特征。实验表明该方法(AOIGen)能够满足大数据量的客户行为分析要求,比其它方法具有直观、高效等特点。The data cube technique is associated with concept hierarchy analysis to integrate attribute-oriented induction （AOI） algorithm and K-means clustering algorithm to form an efficient customer clustering algorithm AOIGen. It is applied to clustering analysis of customer time sequence data to make every sort of customers possesses the similar time sequence character. Experimental results indicated that such integration satisfies the demand of customer behaviour analysis with large amount of data, and has the characters of direct perception and high efficiency compared to other algorithms.

关 键 词：客户行为分析 聚类分析 K-MEANS算法 数据立方体 面向属性的归纳(AOI) 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TN41[自动化与计算机技术—计算机科学与技术]