改进购物篮分析的关联规则挖掘算法  

Association Rule Mining Algorithm on Improved Market Basket Analysis

作　　者：王德兴[1] 胡学钢[1] 刘晓平[1] 王浩[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009

出　　处：《重庆大学学报（自然科学版）》2006年第4期105-107,141,共4页Journal of Chongqing University

基　　金：国家自然科学基金资助项目(60273044);安徽省自然科学基金资助项目(01042201;050420207)

摘　　要：基于改进传统购物篮分析的关联规则挖掘是在数据处理时引入兴趣度加权的思想,将所有交易中同一类商品的交易量进行归一化处理,根据用户领域知识的要求,计算该类商品的兴趣度加权阈值,从而改进传统的购物篮分析,使所挖掘出的关联规则符合实际,同时减少关联规则挖掘的工作量,提高规则挖掘的效率和准确性.Association rules mining based on improved traditional market basket analysis is that we introduce the idea of interest-weighed into data processing, normalize the amount of the same kind of commodity purchased by customers in all transactions. According to the demand of users＇ domain knowledge, we can calculate the interest-weighed threshold, then improve the traditional market basket analysis, association rules are mined much more practical, the workload of association rules mining has been reduced, the efficiency and veracity of association rules mining are improved.

关 键 词：关联规则 APRIORI算法 频繁项目集 数据挖掘 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]