摘要： 在空间数据仓库中，由于物化视图中空间度量的聚集结果需要占用较大的存储空间，因此只能选择部分空间度量进行物化。而现有的物化视图选择算法大部分只是针对视图选择设计的，没有考虑视图中度量的选择。为此，针对空间度量的区域合并操作，提出基于聚类方法的空间度量物化选择算法。把可合并的空间对象组进行聚类，在每个聚类中计算合并组的收益，当选择收益最大的合并组物化后，只需重新计算该类中合并组的收益，即可较大幅度地减少收益计算的开销。通过实验验证了该算法的优越性。
关键词:
空间数据仓库,
空间度量,
物化视图,
聚类方法
Abstract: In spatial data warehouse, the aggregation results of spatial measures in materialized view require substantial storage space. Parts of spatial measures are selected to materialize. And the existing materialized view selection algorithms are mostly designed for view selection. They can not be applied for handling spatial measures. This paper proposes a spatial measures materialized selection algorithm based on cluster method for spatial region merging operation. All merged groups of spatial object are clustered. In each cluster, the algorithm calculates benefit for every merged group. After the best merged group based on the benefit calculation is selected to materialize, the algorithm only recalculates the benefits of merged groups in the cluster which includes materialized group. Overhead of benefit calculation is greatly reduced. Experimental results show the superiority of the algorithm.
Key words:
spatial data warehouse,
spatial measure,
materialized view,
clustering method
中图分类号:
梁银. 基于聚类方法的空间度量物化选择算法[J]. 计算机工程, 2011, 37(8): 58-60.
LIANG Yin. Materialized Selection Algorithm for Spatial Measure Based on Clustering Method[J]. Computer Engineering, 2011, 37(8): 58-60.