Weighted Combination Method of Conflict Evidence Based on Deviation

Abstract:

To solve the problem that D-S evidence combination method may lose effectiveness when evidences are highly or full conflict, we propose weighted combination method of conflict evidence based on deviation. We think, different evidence source is under different circumstance, so the importance of each evidence is different. It is necessary to give weight to every evidence, then conflict probability can only just be assigned. Deviation is a good standard to measure errors. In the circumstance the difference between probabilities from certain evidence source and their mean is too large, that indicates confidence of the evidence is small. So we can use deviation to structure the weight coefficient of each evidence. Based on the weight coefficients, conflict evidences are combined. This method corresponds to people’s thinking habit, only little calculation is needed, and satisfying consequence can be achieved.

• Periodical: Advanced Materials Research (Volumes 457-458)
• Pages: 1581-1585
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.457-458.1581
• Online since: January 2012
• Authors: Jian Liu
• Keywords: Combination, Conflict Evidence, Deviation

References

[1] Zadeh . A simple view of the Dempster-Shafer theory of evidence and its implication for the rule of combination. AI Magazine Vol. 7(2)(1986), p.85.

[2] G Shafer, R Logan . Implementing Dempster's rule for hierarchical evidence. Artificial Intelligence, Vol. 33(1987), p.271.

DOI: 10.1016/0004-3702(87)90040-3

[3] R R Yager . On the Dempster-Shafer framework and new combination rules [J]. Information Sciences, Vol. 41(1987), p.93.

DOI: 10.1016/0020-0255(87)90007-7

[4] D Dubois, HPrade. Represent and combination of uncertainty with belief functions and possibility measures [J]. Comput. Intell ., Vol. 4(1988), p.244.

[5] P Smets. The combination of evidence in the transferable belief model[J]. IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol. 12(5)( 1990), p.447.

DOI: 10.1109/34.55104

[6] Lefevre E, Colot O, Vannoorenberghe P, Belief Functions Combination and Conflict Management, Information Fusion, Vol. 3(2)( 2002), p.149.

DOI: 10.1016/s1566-2535(02)00053-2

[7] DENG Yong, SHI Wenkang, A Modified Combination Rule of Evidence Theory, Journal of Shanghai Jiaotong University, Vol. 37(8)( 2003), p.1275.

[8] WANG Zhuang, HU Weidong, YU Wenxian, ZHUANG Zhaowen, A Combination Rule of Conflict Evidence Based on Proportional Belief Assignment, Acta Electronica Sinica, Vol. 29(12A)( 2001), p.1852.

[9] LI Bi-cheng, WANG Bo, WEI Jun, etc, An Efficient Combination Rule of Evidence Theory, Journal of Data Acquisition & Processing, Vol. 17(1)( 2002), p.33.

[10] YE Qing, WU Xiao-ping, SONG Ye-xin, Evidence combination method based on the weight coefficients and the confliction probability distribution, Systems Engineering and Electronics, Vol. 28(7)( 2006), p.1014.