Abstract
Different fuzzy clustering algorithms, namely hierarchical fuzzy clustering, hierarchical and horizontal samples and characteristics clustering and a new clustering technique, namely fuzzy hierarchical cross-classification were applied to the study of several German and Romanian natural mineral waters using data obtained from samples collected from different sampling sites covering, for example, a large part of the Romanian natural mineral waters. The characteristics clustering technique produces fuzzy partitions of the mineral waters properties involved and thus is a useful tool for studying (dis)similarities between different ions (speciation). The crossclassification algorithm provides not only a fuzzy partition of the mineral waters analysed, but also a fuzzy partition of the characteristics considered. In this way it is possible to identify which ions or other physico-chemical features are responsible for the similarities or differences observed between dierent groups of mineral waters.
*Dedicated to Professor G. E. Baiulescu on the occasion of his 70th anniversary.
Acknowledgments
Notes
*Dedicated to Professor G. E. Baiulescu on the occasion of his 70th anniversary.