一种快速生成最小浓缩数据立方的算法

Fast Algorithm for Computing the Minimal Condensed Cube

语义OLAP技术是近来学者研究的热点之一,浓缩数据立方就是其中一种.本文设计了一个用于快速生成最小浓缩数据立方的算法SQCube.算法分两个阶段:首先利用BottomU pBST算法生成一个非最小的浓缩数据立方,然后对所得到的非最小浓缩数据立方进行后处理,把其中的所有纯BST和隐BST压缩为一条BST,从而生成一个最小浓缩数据立方.实验表明SQCube算法明显优于以往提出的同类算法M inCube.

The semantic OLAP has been widely studied by many researchers recently. This paper designed a fast algorithm, namely SQCube, to compute the minimal condensed cube quickly. The algorithm takes two phases. It first employs the BottomUpBST algorithm to generate a non-minimal condensed cube, and then post-processes it by squeezing all the pure BSTs and hidden BSTs into one BST to generate a minimal condensed cube. Experiments show that SQCube can significantly outperform the MinCube algorithm, which is used for the same purpose.