PROJECTION OF PRIM MINIMAL SPANNING TREE INTO A KOHONEN NEURAL-NETWORK FOR IDENTIFICATION OF AIRBORNE PARTICLE SOURCES BY THEIR MULTIELEMENT TRACE PATTERNS
D. Wienke et Pk. Hopke, PROJECTION OF PRIM MINIMAL SPANNING TREE INTO A KOHONEN NEURAL-NETWORK FOR IDENTIFICATION OF AIRBORNE PARTICLE SOURCES BY THEIR MULTIELEMENT TRACE PATTERNS, Analytica chimica acta, 291(1-2), 1994, pp. 1-18
A hybrid pattern recognition method has been developed as an alternati
ve receptor modelling technique for the identification of sources of c
oarse airborne particles. The Kohonen self-organizing neural network i
s first applied to yield a topological map of an m-dimensional variabl
es space. Unfortunately, during the projection into a low-dimensional
subspace, most of the information about the correct distance between t
he sample vectors is lost. However, the Kohonen network is a useful a
priori step of data compression before application of the minimal span
ning tree. Prim's minimal spanning tree partly compensates for this lo
ss yielding the distance interrelationships between groups of the samp
les. This combination of both projection techniques can overcome some
of their individual deficiencies. Several illustrative examples are pr
esented to demonstrate the nature of the analysis results. Then a set
of airborne particle compositions for samples obtained at a single sam
pling site were analysed. After transferring the combined map to a geo
graphical unit circle (GUC), a correct pattern of the main industrial
emission sources around a sampling site in Granite City (Illinois, USA
) has been obtained by decoding a 35-dimensional space of chemical-ana
lytical variables into a visually and geographically interpretable 2-D
space.