Classification and modeling of different fractions of aerosol monitoring data

Abstract

The present study deals with the application of self-organizing maps (SOM) of Kohonen and four-way Tucker method for the classification and modeling, respectively, of aerosol monitoring data sets from two sampling points (Arnoldstein and Unterloibach) located close to the border, between Austria and Slovenia. The goal of the chemometric data treatment was to find some specific patterns in the classification maps and in the four-way model of complexity [2422] for 5 different aerosol fractions collected in 4 different seasons of the year. The results indicated a distinct separation of the ultrafine particles (PM 0.01–PM 0.4) from the other fractions, which underlines their specific effect on human health. Seasonal separation (but only between summer and winter sampling) is also observed.

Keywords:

• Four-way Tucker method
• SOM
• classification
• modeling
• aerosol
• fraction

Acknowledgements

The authors would like to express their sincere gratitude for the financial support (Project Union DO-02-82), which made it possible to finish this study.