Combining belief functions when evidence conflicts

The use of belief functions to represent and to manipulate uncertainty in e xpert systems has been advocated by some practitioners and researchers. Oth ers have provided examples of counter-intuitive results produced by Dempste r's rule for combining belief functions and have proposed several alternati ves to this rule. This paper presents another problem, the failure to balan ce multiple evidence, then illustrates the proposed solutions and describes their limitations. Of the proposed methods, averaging best solves the norm alization problems, but it does not offer convergence toward certainty, nor a probabilistic basis. To achieve convergence, this research suggests inco rporating average belief into the combining rule. (C) 2000 Elsevier Science B.V. All rights reserved.