The aim of this work is to study the pyramidal clustering of classic and symbolic data. We begin by introduce the different types of variables as well as the notions of compatibility between an order and an index which are necessary to the introduction of pyramids. Next we detail the visualization of pyramids in terms of pyramidal structure. Concepts and properties of indexing and pyramidal index are then presented. Since the set of pyramids is an extension of the set of hierarchies, ties between this two families are developed in order to resolve the problem of saturated pyramids. Finally we give an algorithm of pyramidal clustering provided by the software SODAS 2 and we end by applications of the pyramidal clustering to artificial and real symbolic data.
|Date of Award||2004|
|Supervisor||Andre Hardy (Supervisor), Jean Paul Rasson (Jury) & Pascale Lallemand (Jury)|