一种改进的模糊聚类方法在大气环境质量评价中的应用  

Application of An Adapted Algorithm of Fvzzy Clustering Analysis to the Atmospheric Environmental Quality Assessment

作　　者：张辉军[1] 邓新民[1] 

机构地区：[1]成都气象学院

出　　处：《四川环境》1991年第3期34-40,共7页Sichuan Environment

摘　　要：一般的模糊聚类分析方法只能解决大气污染状况的顺序问题,或只能得出此地与彼地的大气质量状况的相似程度,而不能同时确定大气质量的具体等级。本文提出的一种改进的模糊聚类分析方法解决了这一问题、该方法的主要改进在于:①以各污染因子的污染程度的分级标准值为聚类中心,待分样本与聚类中心之间的相似系数或待分样本隶属于某一类的隶属函数作为聚类函数;②相似系数设计为几何平均最小型,隶属函数设计为正态分布型。实例计算和比较表明,该方法是大气环境质量评价中的一种较好的方法。The general algorithm of fuzzy clustering analysis can only give the similarity of the atmospheric qualities between two regions, but can not determine the specific grade of at mospheric quality To overcome this difficulty, a adapted algorithm of fuzzy clustering analysis is presented in this paper. The algorithm has two obvious points: (1) the grade values of the pollutant level of each pollutant element are considered as the clustering centers, and tire similarity coefficient between the classifying sample and the clustering center or the membership function that the classifying sample belongs to some class is taken as the clustering function;(2) the similarity coefficient is designed as geometrical awerage mini mum model, and the membership function as normal distribution model. After calculating through an example and comparing with other methods, It is showed that the adapted algorithm is an effective method in the at mosph eric envirommental quality assessment.

关 键 词：聚类分析 大气环境 大气质理评价 

分 类 号：X823[环境科学与工程—环境工程]