A CLASSIFICATION OF SEDIMENT TYPES BASED ON STATISTICAL MULTIVARIATE TECHNIQUES

In this paper the results of a characterization and classification of the mineralogical and chemical composition of some rural soils located close to five clay outcrops that supply the Castellon (spanish provin ce) ceramic industry are presented. A total of 50 sediment samples wer e collected in accordance with statistical criteria and their chemical and mineralogical composition were determined. The classification of sediment types has been carried out using complementary statistical te chniques: a hierarchical cluster analysis using Ward method and Euclid ean distance to detect statistical significative groups (clusters) and a discriminant analysis of proximities to obtain probabilities. Univa riate statistical descriptions have been computed for all variables. T he statistical interpretation of the results suggests the existence of three main groups. Comparisons among them are computed.