提取有效规则的关联分类算法  

An Associative Classification Algorithm by Distilling Effective Rules

作　　者：武建华[1,2] 沈钧毅[1] 方加沛[2] 

机构地区：[1]西安交通大学电子与信息工程学院,西安710049 [2]暨南大学珠海学院计算机科学系,广东珠海519070

出　　处：《西安交通大学学报》2009年第4期22-25,共4页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(60673087)

摘　　要：针对关联分类算法产生的规则普遍存在分类器分类精度、效率低的问题,提出了一种提取有效规则的关联分类算法———ACDER算法.首先定义了剩余支持度和剩余置信度,然后通过计算规则剩余支持度和剩余置信度建立了分类器并进行剪枝,以达成对分类尽量少且最有效的规则构成分类器,确保分类器中不存在任何冗余规则和冲突规则.在8个数据集上的测试结果表明,所提算法的平均分类精度比关联规则算法提高了4.15%,而在所有数据源分类器上的规则数却减少了54%.Abstract： Associative classification algorithms commonly have low efficiency and accuracy. A new associative classification algorithm by distilling effective rules, called ACDER, is presented. Both the remaining support and remaining confidence are defined. Then association classifier is constructed and pruned by distilling the most effective rules to ensure that there exists no any re- dundant and conflictive rules in the classifier. Experiment results on eight data sets show that the average accuracy of the classifier is 4.15% higher while the average number of rules in classifier is 54% lower than the CBA classification method.

关 键 词：关联规则 关联分类 有效规则 分类器 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]