ISODATA算法的步骤：（1）预置设定的聚类分析方法…

ISODATA算法 算法步骤： (1) 预置 a. 设定聚类分析控制参数： lc=预期的类数， lNc=初始聚类中心个数(可以不等于c)， lθn =每一类中允许的最少模式数目(若少于此数不能单独成为一类)， lθs =类内各分量分布的距离标准差上界(大于此数就分裂)， lθD=两类中心间的最小距离下界(若小于此数，这两 -ISODATA algorithm algorithm steps: (1) preset settings a. Cluster analysis control parameters: lc = expected number of categories, lNc = initial number of cluster centers (which may not equal to c), lθn = each category of permit the minimum number of models (if less than that and should not become a separate category), lθs = category weight distribution of the standard deviation of the distance between the upper bound of (larger than this number on the split), lθD = two types of the minimum distance between centers of the lower bound of ( If the number is smaller than this, these two