Classification of herbal mixtures on the basis of some metals content using pattern recognition techniques

The purpose of this study was to investigate the content of trace elements in relation to the composition of domestic herbal mixtures. Cluster analysis, principal components analysis, and self-organizing maps were applied to identify existing relationships. The concentrations of Ni, Pb, Cd, Cr, Cu, Fe, Mn, Zn in 81 samples of herbal mixtures were determined with application of FAAS. The study showed that the levels of trace elements in some herbal mixtures of the same composition were comparable. The projection of herbal samples onto topological maps of adjustable sizes allowed recognition of the identical herbal preparations characterized by dissimilar levels of trace elements. The elements which played the most important role in recognition of the herbal samples were Mn, Ni, Zn and partly Cu, Fe and Cd.