Title :
Belief Functions and Probability for General Spaces
Author_Institution :
Dept. of Comput. Sci., San Jose State Univ., San Jose, CA, USA
Abstract :
This paper has two major contributions: (1) Shafer\´s belief functions are extended from finite sets to general universes (sets). (2) Belief function theory on a universe Ω with countable focal elements is interpreted as a probability theory on the product space Ω × [0,1] A counter intuitive point of a belief function is its ignoring the impact of the intersections among focal elements. One obvious intuitive rational is: The basic probability assignment (bpa) is a measurement of some disjoint "information" on focal elements. Item #2 show that this disjoint "information" is a partition (equivalence relation) on the product space. Here are interesting corollary and new exploration: (3) Belief function is an inner probability, if all focal elements are disjoint. (4) Issues for Bel of uncountable focal elements are lightly touched.
Keywords :
belief maintenance; equivalence classes; probability; set theory; basic probability assignment; belief function; equivalence relation; finite sets; focal elements; Area measurement; Calculus; Convergence; Finite element methods; Limiting; Rough sets; Tin; Moore-Smith convergence; belief function; fuzzy set; measure; probability;
Conference_Titel :
Granular Computing (GrC), 2010 IEEE International Conference on
Conference_Location :
San Jose, CA
Print_ISBN :
978-1-4244-7964-1
DOI :
10.1109/GrC.2010.159