基于区间型特征向量的模式识别方法

A pattern recognition method based on interval feature vector

由于客观事物的复杂性和不确定性以及人类思维的模糊性,在模式识别中所要刻画的特征信息往往不是以具体的数值来表示,而是以区间数的形式表示.本文首先基于区间数的中点和宽度以及中点影响因子和宽度影响因子给出区间数距离的计算方法,进一步考虑到诸特征在识别中所占有的重视程度给以赋权,并给出贴近度计算及择近原则.在此基础上,建立模式识别的方法.

Due to the complexities and uncertainties of objective things in the real world,as well as the fuzziness of human thinking,the feature information characterized by pattern recognition often cannot be represented by concrete numerical values,but described by the form of interval numbers. This paper,based on the middle point,width and their impact facts,the method for calculating the distance of interval numbers are presented,and further the weight for every feature is given according to its importance degree in recognition processing. A similarity measure and the nearest neighbor principle are also given. Finally,a new pattern recognition method is established.