一种用于模式识别的多色Voronoi图

提出了一种新的Voronoi图———多色Voronoi图 ,并证明了它的一些重要性质。多色Voronoi图是在Voronoi图的构造过程中 ,根据数据集的特性给Voronoi图染色的结果。用多色Voronoi图概念设计的模式分类器(MCVC)是一种非线性多类别最优分类器 ,可以解决线性可分和线性不可分问题 ,同时MCVC对新增加的训练样本具有快速局部学习能力 ,这种局部学习能力克服了神经网络方法常出现的过学习问题 ,很类似于人对概念的学习。将多色Voronoi图概念用于支撑矢量机的构造中 ,进行支撑矢量的预选取 ,可以提高支撑矢量机的学习速度

Novel Voronoi diagram for pattern recognition

A novel Voronoi diagram, multi-color Voronoi diagram is proposed. Several important features of multi-color Voronoi diagram are proved. Multi-color Voronoi diagram is the coloration result of Voronoi diagram based on the characteristics of data set. Multi-color Voronoi diagram can be used to solve some basic problems in pattern recognition, such as optimal classifier design, MCVC, which can be used to solve linear and nonlinear classification problems. In the meantime, it has the ability to learn locally from new data samples rapidly, which overcomes overfilling problem in neural network method. This ability is similar to that of human being. Multi-color diagram is effective to pre-select the support vector in the construction of support vector machine, which can improve the learning speed of support vector machine.