面向ROLAP的时态垂直粗分区方法  

ROLAP-Oriented Temporal Vertical Partitioning Method Based on Rough Sets

作　　者：李文海[1] 冯玉才[2] 马晓鸣[3] 付铨[2] 胡文斌[1] 

机构地区：[1]武汉大学计算机学院计算机工程系,武汉430079 [2]华中科技大学计算机科学与技术学院数据库与多媒体技术研究所,武汉430074 [3]武汉大学经济与管理学院

出　　处：《计算机学报》2008年第7期1109-1121,共13页Chinese Journal of Computers

基　　金：国家“八六三”高技术研究发展计划项目基金(2004AA4Z3020,2005AA4Z3030);科技部电子政务关键技术及应用系统研究项目(2001BA110B01);湖北省自然科学基金(2006ABA218)资助

摘　　要：数据仓库为海量数据上的决策支持提供了一个高效的信息管理平台,ROLAP利用关系型数据仓库操纵灵活和技术成熟等优势,为面向数据仓库的分析和决策提供了有效的存取、建模和操作方法.然而,传统关系存取方法造成ROLAP的I/O有效性面临严峻的挑战.首先通过分析DSS应用的特点,提出了关系算子访问基表属性的时态行为,定义了算子对属性的时态局部访问.通过对查询样本集的解析建立算子与属性的时态访问映射矩阵,将有效增益作为属性的聚类准则得到时态访问模型PD.最后,给出了求解该模型的粗集算法以及依据聚类结果设计的基表属性的垂直分区方案.实验证明:在决策支持应用中,该方法的效率优于同类的其它优化分区方法.As the essential element of DSS applications, data warehouse coupled with ROLAP provides an administrant platform with high performance. Especially, it exhibits great efficiency in model construction, data access and operation. Conventional access method of the relational data warehouse is, however, faced with new challenges due to the poor I/O efficiency in ROLAP. In this paper, the characters of DSS are analyzed and the temporal-locality attributes accessed by the operators are proposed in terms of the temporal operations. After this, the corresponding mapping matrix of the temporal access is established by parsing the query template set. With the rough set model, the model PD takes the relational attributes for the clustering objects and regards the queries as the vector attributes. With the proposition of rough clustering algorithm, a series of partitions are obtained through a measurement named effectual benefit criteria under different indiscernibility thresholds. At last, the partitions are evaluated and the storage scheme derived from the most effective partition is designed for the vertically partitioning sub-tables of the relations. Experiments show the proposed scheme is superior to other optimization strategies.

关 键 词：关系算子 时态聚类 关系存储 不可辨识度 粗集模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]