摘要
Subjective logic provides a means to describe the trust relationship of the realworld. However, existing fusion operations it offers Weal fused opiniotts equally, which makes it impossible to deal with the weighted opinions effectively. A. Jcsang presents a solution, which combines the discounting operator and the fusion operator to produce the consensus to the problem. In this paper, we prove that this approach is unsuitable to deal with the weighted opinions because it increases the uncertainty of the consensus. To address the problem, we propose two novel fusion operators that are capable of fusing opinions according to the weight of opinion in a fair way, and one of the strengths of them is improving the trust expressiveness of subjective logic. Furthermore, we present the justification on their definitions with the mapping between the evidence space and the opinion space. Comparisons between existing operators and the ones we proposed show the effectiveness of our new fusion operations.
Subjective logic provides a means to describe the trust relationship of the realworld. However, existing fusion operations it offers Weal fused opiniotts equally, which makes it impossible to deal with the weighted opinions effectively. A. Jcsang presents a solution, which combines the discounting operator and the fusion operator to produce the consensus to the problem. In this paper, we prove that this approach is unsuitable to deal with the weighted opinions because it increases the uncertainty of the consensus. To address the problem, we propose two novel fusion operators that are capable of fusing opinions according to the weight of opinion in a fair way, and one of the strengths of them is improving the trust expressiveness of subjective logic. Furthermore, we present the justification on their definitions with the mapping between the evidence space and the opinion space. Comparisons between existing operators and the ones we proposed show the effectiveness of our new fusion operations.
作者
ZHOU Hongwei1,2,3,SHI Wenchang1,2,LIANG Zhaohui1,2,LIANG Bin1,2 1.Key Laboratory of Data Engineering and Knowledge Engineering,Ministry of Education,Beijing 100872,China
2.School of Information,Renmin University of China,Beijing 100872,China
3.Institute of Electronic Technology,Information Engineering University of PLA,Zhengzhou 450004,Henan,China
基金
Supported by the National Natural Science Foundation of China (61070192, 91018008, 60873213, 60703103)
the National High Technology Research Development Program of China (863 Program) (2007AA01Z414)
Natural Science Foundation of Beijing (4082018)
Open Project of Shanghai Key Laboratory of Intelligent Information Processing (IIPL-09-006)
