一种支持结构化P2P的多维范围查找方法

提出一种基于结构化P2P的分层聚类查询系统. 利用空间填充曲线建立从高维特征空间到一维相邻空间的映射,根据映射过程提出一种分层聚类的概念,并将相似的多维数据归入在相同的聚类中,使聚类内的数据具有更大的相似性与更小的值域区间. 模拟结果显示,该方法可减少查询所需的带宽,具有良好的查询准确度和可扩展性.

Structured P2P Search Method to Support Multi-Dimensional Range Queries

A new method called hierarchical clustering queries system is presented. With space filling curve, the system maps data objects in high dimension semantic space into an adjacent range in one dimensional space. According to the mapping process, it reports on a hierarchical clustering method. Similar multi-dimensional data maps to same cluster and the data in same cluster have more comparability and smaller range space. This method can drastically reduce the bandwidth required for query. Simulation results showed that the method has a good query accuracy and scalability.