近年来,数值和名义属性混合数据空间上的轮廓查询仅局限在单个空间上进行,而实际应用中存在对不同子空间轮廓查询的需求.为此,本文结合IPO-tree Search半物化轮廓的方法,定义了半物化轮廓体的概念,提出通过共享子空间轮廓结果集及查询条件计算半物化轮廓体的算法SMS,并设计了存储半物化轮廓体的索引结构NNAS-tree,满足了不同用户高效实时查询任意子空间轮廓的需求.实验比较证明SMS算法效率优于通过逐个单独计算子空间轮廓求轮廓体的传统算法.
In recent years,the study of the skyline query on data space mixed by numeric and nominal attributes is only confined to proceed in a single space.However,there are many skyline queries demand for different sub-space in practical application.In order to satisfy the needs for different sub-space skyline queries,we combine with IPO-tree Search semi-materialization method and propose the concept of semi-materialization skycube.Moreover,algorithm SMS is presented to efficiently calculate the semi-materialization skycube by sharing the sub-space skyline result sets and the query conditions.Meanwhile,an index NNAS-tree,which can meet the needs of users' skyline queries online in any sub-space of data space mixed by numeric and nominal attributes,is designed to store semi-materialization skycube.Finally,comparing with the nave algorithm,which calculates the skycube by separately computing every sub-space skyline,our extensive performance evaluations comfirm the effectiveness of SMS.