R型聚类与模糊聚类分析在源解析中的应用

近年来,聚类分析在环境科学研究中的应用越来越多,R型聚类和模糊聚类是2种常用的聚类分析方法。利用这2种方法对龙岩市大气颗粒物源解析工作中的源样品数据进行聚类分析,并对结果进行比较,研究聚类分析方法在源解析工作中的应用问题。结果表明,聚类分析方法在源解析工作中的应用有实际意义,相比较而言,R型聚类方法的聚类过程比较复杂但是结果精细;模糊聚类计算过程简洁但聚类结果相对粗糙。2种方法都可将研究对象分成若干个类别,但是在得到某一特定个数的聚类结果时所对应的相关系数区间取值不一定相同。最后,讨论了聚类分析在源解析工作中应用的意义和前景。

R-type Clustering Analysis and Fuzzy Clustering Analysis in the Source Apportionment

In recent years,cluster analysis was used more and more in environmental science research.R-typr clustering and fuzzy clustering were two cluster analysis methods commonly used.The above two methods were used for the source samples which were used in the work of source apportionment of atmospheric particles in Longyan City.We compared the results of the two clustering methods to research how we used cluster analysis methods in source apportionment work.The results showed that,cluster analysis method in the application of source apportionment work has practical significance.In comparison,R-clustering process was complicated but the results were fine,Fuzzy clustering process was simple but the results were rough.The research objects could be divided into several categories by the two methods.However when we got specific number of categories of samples,the correlation coefficient of the two methods obtained was not necessarily the same range of values.Finally significance and prospects of cluster analysis used in the application of source apportionment work were discussed.