Statistical biases may be introduced by imprecisely quantifying background radiation reference levels. It is, therefore, imperative to devise a simple, adaptable approach for precisely describing the reference backgro...Statistical biases may be introduced by imprecisely quantifying background radiation reference levels. It is, therefore, imperative to devise a simple, adaptable approach for precisely describing the reference background levels of naturally occurring radionuclides (NOR) in mining sites. As a substitute statistical method, we suggest using Bayesian modeling in this work to examine the spatial distribution of NOR. For naturally occurring gamma-induced radionuclides like 232Th, 40K, and 238U, statistical parameters are inferred using the Markov Chain Monte Carlo (MCMC) method. After obtaining an accurate subsample using bootstrapping, we exclude any possible outliers that fall outside of the Highest Density Interval (HDI). We use MCMC to build a Bayesian model with the resampled data and make predictions about the posterior distribution of radionuclides produced by gamma irradiation. This method offers a strong and dependable way to describe NOR reference background values, which is important for managing and evaluating radiation risks in mining contexts.展开更多
针对传统假设中个体寿命独立同分布的不足,构建了贝叶斯Weibull共享异质性模型,提出了对寿命服从Weibull分布的产品,运用基于Gibbs抽样的马尔可夫链蒙特卡罗(Markov chain Monte Carlo,MCMC)方法动态模拟出参数后验分布的马尔可夫链,在...针对传统假设中个体寿命独立同分布的不足,构建了贝叶斯Weibull共享异质性模型,提出了对寿命服从Weibull分布的产品,运用基于Gibbs抽样的马尔可夫链蒙特卡罗(Markov chain Monte Carlo,MCMC)方法动态模拟出参数后验分布的马尔可夫链,在异质性因子的先验分布为Gamma分布时,给出随机截尾条件下,参数在Weibull共享异质性模型中的贝叶斯估计,提高了计算的精度。借助数据仿真说明了利用WinBUGS(Bayesianinference using Gibbs sampling)软件包进行建模分析的过程,证明了该模型在可靠性应用中的直观性与有效性。展开更多
文摘Statistical biases may be introduced by imprecisely quantifying background radiation reference levels. It is, therefore, imperative to devise a simple, adaptable approach for precisely describing the reference background levels of naturally occurring radionuclides (NOR) in mining sites. As a substitute statistical method, we suggest using Bayesian modeling in this work to examine the spatial distribution of NOR. For naturally occurring gamma-induced radionuclides like 232Th, 40K, and 238U, statistical parameters are inferred using the Markov Chain Monte Carlo (MCMC) method. After obtaining an accurate subsample using bootstrapping, we exclude any possible outliers that fall outside of the Highest Density Interval (HDI). We use MCMC to build a Bayesian model with the resampled data and make predictions about the posterior distribution of radionuclides produced by gamma irradiation. This method offers a strong and dependable way to describe NOR reference background values, which is important for managing and evaluating radiation risks in mining contexts.
文摘针对传统假设中个体寿命独立同分布的不足,构建了贝叶斯Weibull共享异质性模型,提出了对寿命服从Weibull分布的产品,运用基于Gibbs抽样的马尔可夫链蒙特卡罗(Markov chain Monte Carlo,MCMC)方法动态模拟出参数后验分布的马尔可夫链,在异质性因子的先验分布为Gamma分布时,给出随机截尾条件下,参数在Weibull共享异质性模型中的贝叶斯估计,提高了计算的精度。借助数据仿真说明了利用WinBUGS(Bayesianinference using Gibbs sampling)软件包进行建模分析的过程,证明了该模型在可靠性应用中的直观性与有效性。
