Statistics with vague data and the robustness to data representation

Abstract

Robustness is an important problem in statistics. However, robustness of statistical procedures for vague data cannot be limited to insensitivity to departures from assumptions on the underlying distributions. Since the shapes of membership functions applied for modelling vague data are generally strongly subjective one may ask about the influence of these shapes on further decisions.Thus the robustness of the statistical procedures to data representation is also of interest. © 2008 Springer-Verlag Berlin Heidelberg.