Determination of cultural areas based on medieval pottery using an original divisive hierarchical clustering method with geographical constraint (MapClust)

Abstract

Spatial representation based on the distribution of artefact, especially pottery, is widely used in archaeology. The questions raised are often related for the construction of economic or cultural areas, based on fabrics, style or types of vessels. The middle Loire Valley with the study of an important corpus of medieval pottery constitutes an essential framework for understanding this mechanisms over long time. The corpus of data collected and studied for over twenty years includes archaeological sites with reliable chrono-stratigraphic sets whose pottery assemblages are sufficiently large to respond to the problem posed.The permanent increase of the corpus of data, especially since the last publication on this subject in 2013 for which the approach was purely empirical, requires the development of statistical tools adapted here to the spatial clustering analysis of the data. This article deals with a divisive hierarchical clustering method with geographical constraints: MapClust. It is based on spatial indicators called spatial patches. This statistical tool allows the archaeologist to visualize and analyse geographical patterns easily. The comparison of the present results with those of 2013 is a way to see the relevance of the method.