Measuring Conflict Between Possibilistic Uncertain Information Through Belief Function Theory
Dempster Shafer theory of evidence (DS theory) and possibility theory are two main formalisms in modelling and reasoning with uncertain information. These two theories are inter-related as already observed and discussed in many papers (e.g. [DP82, DP88b]). One aspect that is common to the two theories is how to quantitatively measure the degree of conflict (or inconsistency) between pieces of uncertain information. In DS theory, traditionally this is judged by the combined mass value assigned to the emptyset. Recently, two new approaches to measuring the conflict among belief functions are proposed in [JGB01, Liu06]. The former provides a distance-based method to quantify how close a pair of beliefs is while the latter deploys a pair of values to reveal the degree of conflict of two belief functions. On the other hand, in possibility theory, this is done through measuring the degree of inconsistency of merged information. However, this measure is not sufficient when pairs of uncertain information have the same degree of inconsistency. At present, there are no other alternatives that can further differentiate them, except an initiative based on coherence-intervals ([HL05a, HL05b]). In this paper, we investigate how the two new approaches developed in DS theory can be used to measure the conflict among possibilistic uncertain information. We also examine how the reliability of a source can be assessed in order to weaken a source when a conflict arises.
KeywordsBelief Function Possibility Distribution Uncertain Information Possibility Theory Focal Element
Unable to display preview. Download preview PDF.
- [BDP97]Benferhat, S., Dubois, D., Prade, H.: From semantic to syntactic approach to information combination in possibilistic logic. In: Bouchon-Meunier, B. (ed.) Aggregation and Fusion of Imperfect Information, pp. 141–151. Physica Verlag, Heidelberg (1997)Google Scholar
- [DP82]Dubois, D., Prade, H.: On several representations of an uncertain body of evidence. In: Gupta, Sanchez (eds.) Fuzzy Information and Decision Processes, pp. 167–181. North-Holland Publishing Company, Amsterdam (1982)Google Scholar
- [DP87]Dubois, D., Prade, H.: The principle of minimum specificity as a basis for evidential reasoning. In: Bouchon, Yager (eds.) Uncertainty in Knowledge-Based Systems, pp. 75–84. Springer, Heidelberg (1987)Google Scholar
- [DP98b]Dubois, D., Prade, H.: Possibility theory: Qualitative and quantitative aspects. In: Gabbay, Smets (eds.) Handbook of Defeasible Reasoning and Uncertainty Management Systems, vol. 1, pp. 169–226. Kluwer Academic Publisher, Dordrecht (1998)Google Scholar
- [DP01]Dubois, D., Prade, H.: Possibility theory in information fusion. In: Riccia, Lenz, Kruse (eds.) Data Fusion and Perception. CISM Courses and Lectures, vol. 431, pp. 53–76. Springer, Heidelberg (2001)Google Scholar
- [DNP00]Dubois, D., Nguyen, H., Prade, H.: Possibility theory, probability and fuzzy sets, Misunderstandings, bridges and gaps. In: Dubois, Prade (eds.) Fundamentals of Fussy Sets. The Handbooks of Fuzzy Sets Series, Ch. 7, pp. 343–438Google Scholar
- [HL05b]Hunter, A., Liu, W.: A context-dependent algorithm for merging uncertain information in possibility theory (submitted, 2005)Google Scholar
- [HL06]Hunter, A., Liu, W.: Fusion rules for merging uncertain information. Information Fusion Journal 7(1), 97–134 (2006)Google Scholar
- [Liu06]Liu, W.: Analyzing the degree of conflict among belief functions. Artificial Intelligence (in press, 2006)Google Scholar