Abstract | ||
---|---|---|
Dempster-Shafer evidence theory (DST) is an important tool for uncertainty modeling and reasoning, where the uncertainty reasoning includes both the evidence combination and conditioning. New conditioning rules for evidence updating in DST are proposed in this paper. First, two new definitions, namely the weak conditional basic belief assignment (BBA) and strong conditional BBA, are proposed in the spirit of conditional probability in the probabilistic framework. Then, the corresponding Jeffrey-like conditioning rules are proposed to update evidence. The proposed methods have some desirable properties for the evidence updating. Some numerical examples are provided, where existing conditioning rules in DST are compared with newly proposed methods. Experimental results and related analyses show that the conditional BBAs and Jeffrey-like rules proposed in this paper are rational and effective. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1080/03081079.2017.1323891 | INTERNATIONAL JOURNAL OF GENERAL SYSTEMS |
Keywords | Field | DocType |
Belief functions, conditioning, evidence updating, Jeffrey's rule, Bayesian rule | Conditional probability,Basic belief,Conditioning,Artificial intelligence,Uncertainty modeling,Mathematics,Machine learning,Probabilistic framework | Journal |
Volume | Issue | ISSN |
46 | 6 | 0308-1079 |
Citations | PageRank | References |
0 | 0.34 | 20 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Deqiang Han | 1 | 218 | 22.90 |
Yi Yang | 2 | 64 | 5.24 |
Chongzhao Han | 3 | 446 | 71.68 |