基于切片Gibbs抽样算法的空间误差模型的贝叶斯参数估计  

Bayesian parametric estimations for spatial error models based on slice-Gibbs sampling

作　　者：李澳归 赵远英[2,3] LI Aogui;ZHAO Yuanying(College of Mathematics and Statistic,Guizhou University,Guiyang 550025,Guizhou,China;College of Science,Guiyang University,Guiyang 550005,Guizhou,China;Department of Railway Engineering,Guizhou Communications Polytechnic University,Guiyang 551400,Guizhou,China)

机构地区：[1]贵州大学数学与统计学院,贵州贵阳550025 [2]贵阳学院理学院,贵州贵阳550005 [3]贵州交通职业大学轨道交通工程系,贵州贵阳551400

出　　处：《山东大学学报(理学版)》2025年第3期116-126,共11页Journal of Shandong University(Natural Science)

基　　金：国家自然科学基金资助项目(12161014);贵州省省级科技计划项目(黔科合支撑[2023]一般139)2018年度贵州省高层次创新型人才项目阶段性成果。

摘　　要：提出一种切片Gibbs抽样的马尔可夫链蒙特卡罗(Markov chain Monte Carlo,MCMC)算法来计算空间误差模型未知参数的联合贝叶斯估计,通过2个模拟仿真说明提出的贝叶斯估计方法的有效性与切片Gibbs抽样算法的优势,实证分析说明模型和提出的贝叶斯估计方法的有效性。A Markov chain Monte Carlo(MCMC)technique called the slice-Gibbs sampling algorithm is proposed to calculate joints Bayesian estimations of unknown parameters for spatial error models,the proposed algorithm and Bayesian approach are illustrated by two simulation studies,the model and methodology are demonstrated by a real example.

关 键 词：贝叶斯估计 GIBBS抽样 切片抽样 空间误差模型 

分 类 号：O212.8[理学—概率论与数理统计]