基于点击流的客户行为分析引擎的设计与实现  

Design and implement of customer-behavior analyzing engine based on clicks stream

作　　者：杨斌[1] 武彬[2] 

机构地区：[1]湖北工业大学管理学院 [2]华中师范大学图书馆,湖北武汉430079

出　　处：《计算机工程与设计》2007年第13期3042-3045,共4页Computer Engineering and Design

摘　　要：给出了一个基于点击流的客户分析引擎的设计,通过对ExLF日志文件格式进行扩充,分别利用基于嵌入Session变量的Cookie识别机制和时间窗口模型来对用户及用户访问事务进行识别,同时给出了数据源收集的具体数据库表格。在结合聚类和CLV价值分析识别核心客户后,进一步将点击流会话事务文件与商务网站的内部数据相结合,利用改进的Apriori算法,进行多维关联规则挖掘,从而实现对用户行为和兴趣模式进行分析。A design of customer-behavior analyzing engine based on clicks stream is proposed. ExLF log file formats are expanded, and users with cookie recognition mechanism and embedded session variable are identified, User is recognized visit affairs with the time window model, and detailed database tables is presented for the data source. Through clustering and CLV value analysis, core customers is recognized, then, clicks stream dialog files are combined with commercial website＇s inner data, and multidimensional associative rules is mined using improved Apriod algorithm to analyze customer behavior and interests models.

关 键 词：点击流 聚类分析 客户生命周期价值 客户行为挖掘 多维关联规则 APRIORI算法 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]