Representation and Classification of Multidimensional Data Using Self-Organizing Maps
; Enache, G.A.E.
Representation and Classification of Multidimensional Data Using Self-Organizing Maps, Proc Conferência Nacional da Sociedade Portuguesa de Metrologia, Funchal, Portugal, Vol. 1, pp. x1 - x3, October, 2007.
Digital Object Identifier:
Multidimensional data are difficult or even impossible to represent (to visualize) directly. Self-organizing maps (SOMs) are artificial neural
networks conceived by T. Kohonen with the purpose of resolving the problem by converting data of an n-dimensional input space into data of usually a 2D output space. However, in some cases, SOMs can act as a clustering technique and can thus be of further use namely in data mining and for classification purposes. In this paper, and after an introduction where SOMs are described, the authors present examples where
both multidimensional data representation and classification capabilities of SOMs are evidenced.