Probit模型的Bayesian分析及应用  

Bayesian Analysis and Application for the Probit Model

作　　者：朱善维[1] ZHU Shanwei(College of Economics and Management, Shanghai Maritime University, Shanghai 201306, China)

机构地区：[1]上海海事大学经济管理学院,上海浦东201306

出　　处：《钦州学院学报》2017年第1期23-26,共4页Journal of Qinzhou University

摘　　要：针对二值Probit模型中的参数估计问题,采用随机游走M-H算法对Probit模型进行贝叶斯分析,在获得模型参数的后验经验分布的基础上给出参数的具体抽样策略。模拟实验显示:对于中小型样本,在最小后验均方误差准则下获得的参数的估计值接近真实值。对我国2000-2016年外汇实际数据分析表明:目前经济基本面指标表现良好,发生外汇危机的可能性较小。Aiming at the parameters estimation for the Probit model,with aid of the random walk M-H algorithm, the Bayesian analysis was carried out. On the basis of the posterior empirical distribution which was obtained, the specific sampling strategy of parameters was proposed. For medium and small-sized samples, a simulation experiments showed that：under the norm of minimum posterior mean square error, the estimates were close to the true value of them. Analysis of the actual data of China＇ s foreign exchange from 2000 to 2016 showed that the current economic fundamentals indicators were good, and the possibility of foreign exchange crisis was less.

关 键 词：PROBIT模型 Bayesian分析 M-H算法 外汇危机 

分 类 号：O212.8[理学—概率论与数理统计] F830.99[理学—数学]