摘要： 针对超球体查询方式不能将查询范围细化到每一维上的问题，提出一种基于聚类金字塔的超矩形范围查询方法。采用聚类金字塔技术对高维数据建立一维索引值，利用位置保持哈希函数为每个索引值赋予一个唯一的标识，从而构建P-Chord系统，在此基础上，给出范围查询过程。实验结果验证了P-Chord系统在减少距离计算代价与误中点个数、提高查全率等方面的有效性。

关键词: 范围查询, Chord网络, 聚类金字塔, 位置保持哈希函数

Abstract: Conventional super-sphere retrieval method can not set the retrieval scope into every dimension. Aiming at this problem, a retrieval method based on hypercube area is proposed. High-dimensional data are converted into 1-dimensional index values using clustering-pyramid technique and every index value is identified with an unique mark using locality-preserving Hash function. The P-Chord system is generated. A range query process is given on that basis. Experimental results show that the proposed approach has advantages over reducing the cost of computing distance and the sets of false hits, and improving recall ratio.

Key words: range query, Chord network, clustering-pyramid, locality-preserving Hash function

中图分类号: 

• TP311