摘要
本文指出了Dempster合成规则与Zadeh合成规则的局限性和不足.我们建议:对于未确知性信息的评价,如果专家们掌握证据之间的联系,他们就应共享所有证据以便得出合理的结论.这时,可以利用简便的加权平均法.如果专家们难以搞清证据之间的联系,这些证据就该按照其性质分成几组,再利用多级加权平均的方法.这种方法不仅使用方便,而且比现有的信息合成方法更合理.
In this paper,the limitations and weaknesses of Dempster's rule and Zadeh'srule of combination are pointed out.It is suggested,that in the evaluation ofunascertained information,if the experts know the connection between the items ofevidence,they should share all the evidence in order to obtain a reasonable conclu-sion.In this case,the simple weighted mean method can be used.If it's difficultfor the experts to make clear the connection between the items of evidence,the itemsof evidence should be classified into several groups according to their charac-teristics.In this case,we at first use the weighted mean method for each group,and then use the same method for all groups.This way is not only very convenientto use,but also more reasonable than the existing methods of information combina-tion.
关键词
未确知信息
主观概率
信比分布
unascertained information
Dempster's rule
Zadeh's rule
subjective probability
faith ratio distribution
