摘要
基于贝叶斯信息融合与统计推断原理,建立不确定度动态评定模型,对测量不确定度进行实时更新。引入最大熵原理和爬山搜索优化算法,确定先验分布概率密度函数及样本信息似然函数,结合贝叶斯公式求出后验分布概率密度函数,实现不确定度的优化估计。仿真及实例分析表明,基于贝叶斯和最大熵方法评定及更新的测量不确定度更加接近理论值。
Based on the principle of Bayesian information fusion and statistical inference, the dynamic evaluation model of uncertainty was established and the uncertainty of measurement results was updated in real time. The maximum entropy principle and hill-climbing algorithm were introduced to determine the prior distribution probability density function and the likelihood function of the sample information. The distribution of posterior distribution of PDF was calculated by combining the Bayes formula. And the optimal estimation of uncertainty was achieved. The case and simulation showed that the measurement uncertainty obtained by Bias and maximum entropy method was more accord with the standard requirement.
出处
《计量学报》
CSCD
北大核心
2017年第1期123-126,共4页
Acta Metrologica Sinica
基金
国家自然科学基金(51275148)
合肥工业大学青年教师创新项目(JZ2014HGQC0126)
关键词
计量学
不确定度评定
贝叶斯
信息融合
最大熵原理
爬山算法
metrology
uncertainty evaluation
Bayes
information fusion
maximum entropy principle
hill-climbing algorithm
