摘要
从数据点的空间排列特点出发提出了k近邻搜索算法,利用多向链表对数据集进行排序,综合考虑了数据集的范围、点的总数、搜索步长及最近点数目k,并采用了空间包围策略,可以给出接近于最佳搜索速度的步长e和k值,并且在搜索终止准则上进行改进,使近邻点的搜索范围大大缩小,搜索速度加快。
An algorithm to finding the k-nearest neighbors of points is provided quickly which is based on the space character of data-points. Data set is queued by multilinked list. By using envelopment-space, the range of data set, the total numbers of points, the searching step and the numbers of nearest neighbors, the method is easy to obtain the value of e and k for a nearly optimal searching. By improving the terminate rule of searching, this method has some excellent characters such as shorten searching range and quicken speed.
出处
《湖南工业大学学报》
2007年第2期84-87,共4页
Journal of Hunan University of Technology
关键词
K近邻
多向链表
快速搜索
k-nearest neighbors
multilinked list
quickly search
