模糊c-均值算法是模糊聚类中最常见、应用最广泛的分析方法之一．经典FCM算法理论和应用的研究已经相当成熟，Bezdek业已证明算法的收敛性［6］，许多软件提供了多种方便的FCM工具箱(如Matlab中FCM工具箱等)．但是传统FCM算法处理的是普通数据集，不能直接处理区间数，得到的聚类原型也不是区间数．针对区间数，一种直觉的方法是分别对左区间值和右区间值作FCM，并把得到的聚类原型分别作为区间型聚类原型的左右区间值，但这种方法已经被证明是行不通的［6］-fuzzy c-means clustering algorithm is fuzzy most common use of the most extensive analysis of one of the methods. FCM classical theory and application of the research is already quite mature, PET has proved the convergence of the algorithm [6] Many software provides a variety of convenient FCM toolbox (eg Matlab toolbox FCM). But traditional FCM is the common data sets, not directly deal with the interval, The prototype is not clustering interval. Against interval, an intuitive method is to separate the left and right interval interval for the FCM, and the prototype were cluster-cluster as a prototype interval around the interval, But this method has been proven to be unworkable [6]