Métodos geométricos de la estadística

Abstract

Se exponen diversos métodos geométricos, insistiendo en qué propiedades se fundamentan, en orden a probar que diferentes espacios geométricos (euclídeo, no euclídeo, ultramétrico, aditivo, riemanniano) juegan un importante papel en el análisis estadístico de datos.

In this paper we discuss the important role of the geometric approach in statistical data analysis. We show as Borne geometric structures, as euclidean, non-euclidean, ultrametric, additive and riemannian spaces, may be considered as the theoretical foundations of some statistical criteria and proximity data analysis. We specially study the representation of a set with a statistical distance on a geometric model space.