基于混合型模糊聚类分析的降水区域分类法  

Precipitation Regional Classification Based on Mixed Fuzzy Clustering Analysis

作　　者：邹杰涛[1] 汪海燕[1] 赵方霞[1] 

机构地区：[1]北方工业大学理学院,北京100144

出　　处：《数学的实践与认识》2015年第2期106-112,共7页Mathematics in Practice and Theory

基　　金：北方工业大学教育教学改革和课程建设基金

摘　　要：在混合型模糊聚类分析的基础上,先用传递闭包法得出动态聚类结果,然后引用F统计量找到最佳分类,针对此最佳分类给出具体算法求得初始分类矩阵,然后利用模糊C-均值算法对初始分类矩阵进行迭代计算,对原分类结果进行软划分修正从而得出最终聚类结果.方法减少了两次人为因素对聚类结果的干扰.将其应用于降水区域划分的实证分析表明,可以对降水区域进行更为有效的划分.Based on mixed fuzzy clustering analysis method, a dynamic clustering result was firstly performed by use of the transitive closure method. According to the result, F statistic was introduced to find a best classification. With the best classification, an algorithm was presented to count an initial partition matrix, then made use of the F^zzy C- means algorithm to iterative calculate to revise the initial partition matrix, consequently get the ultimate clustering result. This method avoided disturb because of two human factors with the result. The effect of application on precipitation regional classification shows, this method can give a more effectively precipitation regional classification.

关 键 词：模糊聚类分析 传递闭包法 模糊C-均值 F统计量 

分 类 号：O212.1[理学—概率论与数理统计]