基于关系代数的关联规则挖掘算法  

Association rule mining algorithm based on relation Algebra theory

作　　者：陈莉[1] 焦李成[2] 

机构地区：[1]西北大学信息科学与技术学院,陕西西安710069 [2]西安电子科技大学雷达信号处理国家重点实验室,陕西西安710071

出　　处：《西北大学学报（自然科学版）》2005年第6期691-694,698,共5页Journal of Northwest University（Natural Science Edition）

基　　金：国家"863"计划基金资助项目(863-306-06-06);陕西省自然科学基金(98X11);陕西省教育厅重点科研计划基金资助项目(00JK015)

摘　　要：目的提出基于关系代数理论的关联规则挖掘算法。方法利用数据预处理方法,剔除无关属性、获得相应的目标特征子集。结果基于目标特征子集,利用关系矩阵及相关运算给出了搜索大项集的基于关系代数理论的优化的关联规则挖掘算法,该算法只需扫描数据库一次。结论克服了经典的Apriori算法需要多次扫描数据库的缺点,同时算法具有良好的并行性和可伸缩性。Aim To put forward association rule mining algorithm based on relation algebra theory. Methods With samples of different object feature subsets usually possess similar pattern; data preprocessing is used to eliminate useless attributes and obtain subsets correspond to object feature. Results An optimized algorithm about association rule mining is proposed, which uses relation matrix and relation algebra operations to search large items base on relation theory. Algorithm searches database only one time. Conclusion The disadvantage of classical apriori algorithm is overcome, such as scan database repeatedly. At the same time the algorithm possesses favorable parallel and scalable explained by experiments.

关 键 词：关系代数 关联规则 支持度 并行挖掘 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]