论文部分内容阅读
提出基于模式训练的Voronoi划分向量分类器结构,指出该结构是以欧几里得距离为测量的最优模式分类器,给出M类k维(k≥2)Voronoi域分类器的两个判决准则和三种结构。
The structure of Voronoi partitioning vector classifier based on pattern training is proposed. It is pointed out that this structure is the optimal pattern classifier based on Euclidean distance and gives two decisions of class M k-dimensional (k≥2) Voronoi domain classifier Guidelines and three structures.