Approximation of belief functions by minimizing euclidean dis-tances

Authors Thomas Weiler
Ulrich Bodenhofer
Editors P. Grzegorzewski
O. Hryniewicz
M. Á. Gil
Title Approximation of belief functions by minimizing euclidean dis-tances
Booktitle Soft Methods in Probability, Statistics and Data Analysis
Type in collection
Publisher Physica-Verlag
Series Advances in Soft Computing
ISBN 3-7908-1526-8
Year 2002
Pages 170-177
SCCH ID# 262
Abstract

This paper addresses the approximation of belief functions by minimizing the Euclidean distance to a given belief function in the set of probability functions. The special case of Dempster-Shafer belief functions is considered in particular detail. It turns out that, in this case, an explicit solution by means of a projective transformation can be given. Furthermore, we also consider more general concepts of belief. We state that the approximation by means of minimizing the Euclidean distance, unlike other methods that are restricted to Dempster-Shafer belief, works as well. However, the projective transformation formula cannot necessarily be applied in these more general settings.