特征指标信息不完全的系统聚类方法

A System Clustering Method with Incomplete Attribute Information

针对聚类对象特征指标值信息不完全且指标权重确定的多指标聚类分析问题,提出了一种新的系统聚类分析方法。在本文中,首先对特种指标值信息不完全的多指标聚类问题进行了描述;然后依据传统的系统聚类分析方法的基本思路,给出了求解特征指标值信息不完全的多指标聚类问题的计算步骤,其核心是通过构建并求解二次规划模型,将不完全信息转化为完全的数值信息的形式,进而通过运用系统聚类法的思路,即可得到所有聚类对象的分类结果。最后通过给出了一个算例说明本文提出的方法。

With respect to multiple attribute clustering analysis problems with incomplete information, in which attribute values are in the form of incomplete information and weight is exactly numerical data, a system clustering analysis method is proposed. In this paper, firstly, the multiple criteria clustering problem with incomplete information is introduced. Then, based on the traditional system clustering analysis method, calculation steps for solving multiple attribute clustering analysis problems with incomplete information are developed. In the method, a quadratic programming model is constructed for changing the incomplete information to exactly numerical data. According to the system clustering analysis method, the clustering result is gotten. Finally, a numerical example is used to illustrate the use of the proposed method.