MULTIVARIATE STATISTICAL INTERPRETATION OF A DATA SET OF MEDIEVAL GLASS FRAGMENTS

The present study deals with the application of cluster analysis and principal components analysis (PCA) to a data set of 25 medieval glass fragments (21 of them excavated on the territory of present Bulgaria). The glasses have been analysed using different analytical techniques. The data obtained for the element composition of the glasses were treated for 7 chemical parameters (SiO2, Na2O, K2O, MgO, CaO, Al2O3 and Fe2O3) by means of z-transformation of the raw data, Ward's method of linkage and squared Euclidean distance as similarity measure in order to discover appropriate linkage between the glass fragments and between the chemical parameters characterizing them. Thus, the investigated glass artefacts are classified into several patterns depending on their chemical composition and are related to the specific characteristics of the raw materials used for their production and/or the glassmaking technology.