1. 湖南涉外经济学院 管理学院,长沙,410205
2. 国防科学技术大学 信息系统与管理学院,长沙,410073
纸质出版:2017
移动端阅览
刘琦, 李旭, 张安扬. 命中精度Bayes推断中基于P值的验前可信度建模[J]. 航空动力学报, 2017, 32(1): 75-81.
doi: 10.13224/j.cnki.jasp.2017.01.011
刘琦, 李旭, 张安扬. 命中精度Bayes推断中基于P值的验前可信度建模[J]. 航空动力学报, 2017, 32(1): 75-81. DOI: 10.13224/j.cnki.jasp.2017.01.011.
doi: 10.13224/j.cnki.jasp.2017.01.011 DOI:
在对散布参数验前可信度的定义与必要性进行分析的基础上,给出了给定散布参数(方差)的P值计算模型,构建了一维散布分析时验前分布可信度的P值模型,给出了验前P值与验前可信度的定义,并对其变动情况进行了阐述,给出了验前分布的拒绝条件.对二维散布分析条件下的验前可信度分析,分别构建了验前P值的独立求解模型和集成求解模型.并从0.05,0.2,0.4区分的数据区间,给出了验前分布选择的原则和要求.基于案例比较可见,验前可信度能较为显著地完成验前分布的差异化比较,相对现有的仅对验前可信度的定性说明,基于P值的验前可信度建模与计算,不仅能完成验前可信度的量化计算,而且能从概率的角度对验前分布选择的正确性进行假设检验.
Based on the analysis of the concept and necessity of prior reliability for dispersion parameters prior distribution
the P value calculating model for given dispersion parameter (variance) was presented
the prior distributions reliability calculating model for onedimensional dispersion parameter was constructed
and the concepts of prior P value and prior reliability were presented. The variation of prior reliability was described
and the reject rule of prior distribution was given. To the analysis of prior reliability for twodimensional dispersion parameter
the independent solution model and integrated solution model for prior P value were carried out respectively. And with the interval divided by 005
02 and 04
the principle and requirement of prior distribution selection was presented. From cases comparison result
it was easy to see that prior reliability can be easy to complete the difference comparison of prior distributions
and compared with traditional method only could describe prior reliability from qualitative view
the P value model not only calculated the numeric of prior reliability
but also could complete the correctness hypothesis test of prior distribution form probability view.
0
浏览量
443
下载量
CSCD
关联资源
相关文章
相关作者
相关机构