导　　师： 吴根秀

学科专业： G0104

授予学位： 硕士

作　　者： ;

机构地区： 江西师范大学

摘　　要： 该文主要提出了两种新的分类方法:基于证据推理模型的k- nn分类方法及基于可变精度粗集模型的k-nn分类方法.在前一种分类方法中,分类专家对待分类样本点的最近邻样本点给出权重,从而定义关键样本点及非关键样本点,进而给出它们的支持度、折扣系数.通过上述概念的引入,可对基于证据理论的k- nn分类方法进行修正,使分类结果更加精确.并且当折扣系数为1,且给定所有最近邻样本点权重相等时,基于证据推理模型的k-nn分类方法就成为基于证据理论的k- nn分类方法.并且给出了例子且进行了计算机模拟实验,取得了较好的效果.后一种分类方法是将可变精度粗集模型与k- nn分类结合起来,从而可通过给定最大容忍的错误分类率来控制分类的准确度,使分类结果达到所期望的目的.并且给出了一些例子. This dissertation mainly presents two new classification methods:the k-NN classification method based on a evidence reasoning model and the k-NN classification method based on variable precision rough set model. During the former classification method,the classification expert gives the weights of the nearest neighbor sample points of the sample point to be classified,then defines the key sample point and non-key sample point,furthermore gives their support degree,discount coefficient. Through the allusion of the former notions,we can ameliorate the k-NN classification method based on evidence theory,and make the classification result more precise. When the discount coefficient is 1 and all weights of the nearest neighbor sample points are the same,the k-NN classification method based on evidence reasoning model will become the k-NN classification method based on evidence theory. Furthermore,a example is given and a computer simulation is performed,then a good result is got. The latter classification method combined variable precision rough set model with k-NN classification method,so we can control the classification accuracy rate by the most endurable given error classification rate,then we can make the classification result conform to what we expect,and also some examples are given.

关 键 词： 分类 证据理论 证据推理模型 粗集 可变精度粗集模型

分 类 号： [O235]

领　　域： [理学] [理学]