多源验前信息下先验分布的稳健融合方法

THE ROBUST FUSION METHOD FOR PRIOR DISTRIBUTION IN MULTI-SOURCES PRIOR INFORMATION

根据多种先验分布与似然函数尾部特性的比较,给出了多源验前信息下先验分布的稳健融合方法.讨论了由该方法得到的融合先验分布的后验稳健性问题.最后的数值例子表明,由该方法得到的融合先验分布具有较好的稳健性,进一步验证了该方法的有效性.

By comparing the tail behavior of prior distributions and likelihood distribution, a robust fusion method of priors distribution in multi-sources prior information is given. The posterior robustness of the fusion prior distribution based on this method is also considered. A numerical example shows that the prior distribution fused by this method is robust, and hence that the method proposed in this paper is effective.