一种基于模糊聚类的多准则分类方法.pdfVIP

一种基于模糊聚类的多准则分类方法 王坚强 （中南大学商学院，湖南长沙 410083 ） 摘 要：针对权系数信息不完全确定且有训练集的多准则分类决策问题，提出了一种基于模糊聚类的分 类方法。该方法在考虑对训练集分类的基础上，结合不完全确定的准则权系数信息等建立模糊聚类模型，通 过粒子群算法求解所得优化模型，得出准则权系数和聚类中心及聚类方案在准则下的效用值，通过插值得到 非聚类方案在准则下的效用值，并计算方案属于各类别的隶属度，进而得到整个方案集的分类。最后实例说 明该方法的有效性和可行性。 关键词：多准则分类决策；模糊聚类；信息不完全确定；粒子群算法 中图分类号：C934 文献标识码：A A Multi-criteria Classification Approach with Incomplete Certain Information Base on Fuzzy Clustering Wang Jian-qiang （School of Business ， Central South University ，Changsha，410083 ，China） Abstract: For multi-criteria classification problems with reference set and the incomplete certain information on the criteria’s weights, a method based on fuzzy clustering is proposed. Using classification to the reference set and incomplete certain information on weights, a fuzzy clustering model is developed. Then using particle swarm optimization algorithms to solve the fuzzy clustering model, the optimal criteria’s weights and clustering center and utility of clustered alternatives under criteria are gained. The utility of non-clustered alternatives under criteria is gained by using interpolation. The membership degree of the alternatives to categories is computed. And classification is performed through the comparison of membership degree of the alternatives to categories. Finally, an example is given to explain the feasibility and availability of this method. Keywords: multi-criteria classification decision-making; fuzzy clustering; incomplete certain information; particle swarm optimization algorithms 1 引 言 分类问题是现实生活和理论研究中重要的一类多准则决策问题。目前，已有一些多准则分类决策 [1~2] [3~4] 方法 ，如ELECTRE TRI、U