摘要
在d维空间n个数据点中,k近邻搜索用于查找给定查询点的k个最近邻居.针对k最近邻搜索算法存在的问题,提出了一种基于P2P的k最近邻自适应搜索算法PKA.探讨了度量空间、相似性查询和GHT*规则,定义了高维数据的相似度函数ESF(X,Y),论述了GHT*中插入算法及范围查找算法和搜索算法.在此基础上,具体给出了PKA算法的实现方法,并验证了其正确性.
Given n data points in d-dimensional space, k nearest neighbors searching involves determining k nearest of these data points to a given query point. An adaptive distributed k-nearest neighbor search algorithm based on P2P called PKA in high dimensions is proposed to solve the shortcomings of KNNs. Metric Space, Similarity Queries and Principles of GHT＊ are discussed. Similarity measure function ESF（X, Y） is given. Insert, Range find and Search Algorithms in GHT＊ are discussed. The detailed PKA algorithm is given and discussed with experiment.
出处
《微电子学与计算机》
CSCD
北大核心
2009年第9期61-63,共3页
Microelectronics & Computer
基金
国家社会科学基金项目(09BJY106)
