Confidence Factors, Empiricism and the Dempster-Shafer Theory of Evidence

ABSTRACT The issue of confidence factors in Knowledge Based Systems has become increasingly important and Dempster-Shafer (DS) theory has become increasingly popular as a basis for these factors. This paper discusses the need for an empirical interpretation of any theory of confidence factors applied to Knowledge Based Systems and describes an empirical interpretation of DS theory suggesting that the theory-has been seriously misinterpreted. For the essentially syntactic DS theory, the empirical model developed is based on the semantics of sample spaces. This model is used to show that, if belief functions are based on reasonably accurate sampling or observation of a sample space, then the beliefs and upper probabilities as computed according to OS theory cannot be interpreted as frequency ratios. Since a number of proposed applications of OS theory use belief functions in situations with statistically derived evidence and seem to appeal to statistical intuition to provide an interpretation of the results, it is likely that OS theory has often been misapplied. CONFIDENCE FACTORS, EMPIRICISM AND THE DEMPSTER-SHAFER THEORY OF EVIDENCE The issue of confidence factors in Knowledge Based Systems has become increasingly important and Dempster-Shafer (DS) theory has become increasingly popular as a basis for these factors.