基于聚类分析的案例距离决策模型

The Study of Case-based Distance Decision Model Based on Clustering Analysis

为解决偏好信息直接获取困难并且评价数据量较大的多属性决策问题,本文提出一种用聚类分析和案例距离相结合的方法,先用聚类方法提取出部分具有代表性数据,构建基于距离的最优化参数拟合模型,再将求解获得的参数应用到全体数据评价,同时定义了三类帕累托约束约简条件,提升了模型的计算效率,最后用研究生创新能力评价算例演示了方法的可用性。

In this paper,a method combining clustering analysis and case-based distance approach is proposed to solve multiple criteria ranking problem.The features of the proposed method contain the simple and easy preference interaction acquisition and efficient way to handle a large data set of alternatives' ranking problem.Firstly,the cluster analysis is designed to identify representative case set and then a casebased distance model is proposed to solve ranking decision analysis.Finally,a case study to evaluate the graduate students' learning ability is carried out to demonstrate the feasibility of the method.