基于粗集的模糊聚类方法和结果评估

粗集的决策表的属性包括定量属性和定性属性，针对这种情况，根据一种对象的相似性度量方法，使用模糊聚类方法对粗集对象进行模糊聚类，对聚类结果进行了评估(根据这种聚类方法得到的结果和实际的分类结果进行比较)．在这种相似性度量方法基础上，证明了粗集的等价关系可以被转化为模糊等价矩阵．基于粗集的聚类步骤如下：首先，一个粗集等价关系都可以转化为一个模糊相似矩阵，其次，转化成一个模糊等价矩阵，最后，进行模糊聚类．对此方法进行了实验，并对实验的结果进行评估．实验结果说明了这种方法的简单高效．

Fuzzy Cluster Based on Rough Set and Result Evaluating

The attributes in decision table of rough set include quantitative attributes and qualitative ones. According to a method of similarity measure, it utilizes fuzzy cluster for the objects of rough set. Then the evaluation of results of cluster is given according to the comparison between what is from the result of cluster and what is from actual classification. Based on the method of similarity measure, it proves that the equivalence relation of rough set can be transformed into fuzzy equivalence matrix completely. The basic steps of cluster is given as following: 1)transforming into a fuzzy similarity matrix from equivalence relation of rough set; 2) forming a fuzzy equivalence matrix from fuzzy similarity matrix; 3)fuzzy cluster. Some experiments and evaluations of the results of experiments are made. The results of these experiments show this method is simple and efficient.