基于子类划分的分类器设计方法

A Classifier Design Method Based on Sub class Division

该文给出了一种基于子类划分的分类器设计方法，提出了一个基于类内散布矩阵和类间散布矩阵的可分性新准则，并提出了利用遗传算法通过训练自动确定子类个数的方法。人脸识别实验表明，所提出的方法能够提高近邻分类方法的效果，并解决学习矢量量化分类方法典型样本个数难以合理确定的问题。

This paper gives a classifier design method based on sub class division, brings forward a new divisibility rule based on the within class scatter matrix and the between class scatter matrix and a method to ascertain the number of sub class by training with the genetic algorithm. Experiments on a face database show that the new classifier design method can improve the classification effect of the neighbour method, and can solve the difficult problem of ascertaining justly the typical samples in the classification of the learning vector quantization.