基于Gibbs抽样方法的空间滞后随机前沿模型Bayesian估计  

Bayesian Estimation of Spatial Lag Stochastic Frontier Model Based on Gibbs Sampling Method

作　　者：窦剑军 王媛 张辉国[2] 胡锡健[2] Dou Jianjun;Wang Yuan;Zhang Huiguo;Hu Xijian(Longqiao College, Lanzhou University of Finance and Economics, Lanzhou 730101, China;College of Mathematics and System Science, Xinjiang University, Urumqi 830046, China)

机构地区：[1]兰州财经大学陇桥学院,兰州730101 [2]新疆大学数学与系统科学学院,乌鲁木齐830046

出　　处：《统计与决策》2018年第7期14-17,共4页Statistics & Decision

基　　金：国家自然科学基金资助项目(41261087);新疆自然科学基金资助项目(2015211C276);新疆大学博士启动基金资助项目(BS130103)

摘　　要：文章应用Gibbs抽样方法对空间滞后随机前沿模型参数进行Bayesian推断,得到模型参数的后验条件分布。应用Gibbs抽样方法对模型参数的后验均值进行了推断,该方法避免了对复杂表达式的高维积分计算。通过蒙特卡罗模拟显示在最小后验均方误差准则下得到的参数估计值十分逼近真值。This paper applies Gibbs sampling method to conduct Bayesian inference on spatial lag stochastic frontier model, and obtain parameters＇ posterior conditional distribution. And then the paper uses Gibbs sampling method to deduce posterior mean of the model parameter. This method avoids the numerical calculation of the complex expression. A Monte Carlo simulation shows that the Bayesian estimate of parameter under the criterion of the minimum posterior mean square error is very close to its true value.

关 键 词：GIBBS抽样 空间滞后随机前沿模型 贝叶斯估计 技术效率估计 

分 类 号：O212[理学—概率论与数理统计]