语义Cube的层次聚类方法

在Data Cube语义特性及模式中维的层次性基础上提出了多维数据之间的等价关系≡-{HCov}，对Data Cube进行层次聚类.该方法的优点在于用等价类的方法保存了所有聚集记录，同时定义了等价类之间的关系，以保存聚集记录之间的层次信息.理论分析和实验证明，该方法不但节省了存储空间，而且利用聚类信息及层次信息，可高效的进行各种OLAP查询，以及支持多维数据上的上钻下钻、旋转等Cube语义操作.同时，为OLAP查询导航、OLAP查询行为分析的实现提供了可能.

A Hierarchical Clustering Method on Semantic Cube

An equivalence relation ≡-{HCov} is proposed based on the Data Cube's semantics and the dimension's hierarchy of its pattern, and then on the basis of the equivalence relation, proceeds the data cube to hierarchical cluster. The advantage of this method not only refers to the preservation of all the aggregate records by using equivalence class, but also means the definition of the classified information as well as hierarchical information. The result of theoretical analysis and experiments indicates that this method can effectively save the storage space, and with clustering information and hierarchical information, it can also provide kinds of OLAP query with high efficiency. Meanwhile, the method has the ability to support some semantic operation such as roll-up/drill-down and rotate, and makes it possible to realize the OLAP query navigation and behavior analysis of OLAP query.