Elucidating the composition of atmospheric aerosols through the combined hierarchical, non-hierarchical and fuzzy clustering of large electron probe microanalysis data sets

Combined utilization of hierarchical, non-hierarchical and fuzzy clustering is introduced to find the intrinsic Structure of large analytical data sets. Modified algorithms for nonhierarchical and fuzzy clustering are used. The interpretation of the results of clustering is based on the extensive use of different types of graphs. New approaches to the graphical representation of the results of fuzzy clustering are proposed. The method is applied to data of electron probe X-ray microanalysis of individual aerosol particles. The advantages of the method are shown.