SPECIALIZATION—A NEW CONCEPT FOR UNCERTAINTY HANDLING WITH BELIEF FUNCTIONS

Abstract Whenever methods for the treatment of uncertainty based on the theory of belief functions are presented in the literature they mostly rely on Dempster's rule of combination. The concept of specialization generalizes this rule and is less restrictive. It is founded on the fact that modelling uncertain phenomena always entails a simplifying coarsening which arises from the renounciation of a description to that depth a perfect image would require.