The modified Dempster-Shafer approach to classification

This paper describes "modified Dempster-Shafer" (MDS), an approach to object identification which incorporates Bayesian prior distributions into an altered Dempster-Shafer rule of combination. The MDS combination rule reduces, under strong independence assumptions, to a special case of Bayes' rule. We show that MDS has rigorous probabilistic foundations in the theory of random sets. We also demonstrate close relationships between MDS and Smets' "pignistic" probabilities (1990), which in the MDS framework become true posterior distributions. We describe the application of MDS to a practical classification algorithm which uses an information-theoretic technique to limit the combinatorial explosion of evidence. We also define a non-ad hoc, MDS-based classification "miss distance" metric used to measure the performance of this algorithm.

[1]  Philippe Smets,et al.  Constructing the Pignistic Probability Function in a Context of Uncertainty , 1989, UAI.

[2]  G. Klir.  Where do we stand on measures of uncertainty, ambiguity, fuzziness, and the like? , 1987 .

[3]  Didier Dubois,et al.  Consonant approximations of belief functions , 1990, Int. J. Approx. Reason..

[4]  D. Dubois,et al.  Properties of measures of information in evidence and possibility theories , 1987 .

[5]  Philippe Smets,et al.  The Transferable Belief Model , 1994, Artif. Intell..

[6]  Ronald P. S. Mahler.  Combining ambiguous evidence with respect to ambiguous a priori knowledge. I. Boolean logic , 1996, IEEE Trans. Syst. Man Cybern. Part A.

[7]  Rodney W. Johnson.  Independence and Bayesian Updating Methods* *Reprinted from the Artificial Intelligence Journal with permission of the publisher, North Holland, Amsterdam. , 1985, UAI 1985.

[8]  Harry E. Stephanou,et al.  Measuring Consensus Effectiveness by a Generalized Entropy Criterion , 1988, IEEE Trans. Pattern Anal. Mach. Intell..

[9]  Ronald P. S. Mahler.  Classification when a priori evidence is ambiguous , 1994, Defense, Security, and Sensing.

[10]  Nic Wilson,et al.  Decision-Making with Belief Functions and Pignistic Probabilities , 1993, ECSQARU.

[11]  Ronald P. S. Mahler,et al.  Combining ambiguous evidence with respect to ambiguous a priori knowledge. Part II: Fuzzy logic , 1995, Fuzzy Sets Syst..

[12]  John Yen,et al.  A Reasoning Model Based on an Extended Dempster-Shafer Theory , 1986, AAAI.

[13]  John Yen,et al.  GERTIS: a Dempster-Shafer approach to diagnosing hierarchical hypotheses , 1989, CACM.

[14]  D. Kerridge.  Inaccuracy and Inference , 1961 .

[15]  Frans Voorbraak,et al.  On the Justification of Dempster's Rule of Combination , 1988, Artif. Intell..

[16]  Philippe Smets,et al.  The transferable belief model and random sets , 1992, Int. J. Intell. Syst..

[17]  Rudolf Kruse,et al.  Uncertainty and Vagueness in Knowledge Based Systems , 1991, Artificial Intelligence.

[18]  Hung T. Nguyen,et al.  On Random Sets and Belief Functions , 1978, Classic Works of the Dempster-Shafer Theory of Belief Functions.