基于改进BVAR模型和MS-VECM模型的能源消费分析  

Energy Consumption Analysis Based on Improved BVAR Model and MS-VECM Model

作　　者：王星 WANG Xing(School of Mathematics and Statistics,Chongqing Technology and Business University,Chongqing 400067,China)

机构地区：[1]重庆工商大学数学与统计学院,重庆400067

出　　处：《重庆工商大学学报（自然科学版）》2023年第6期111-118,共8页Journal of Chongqing Technology and Business University:Natural Science Edition

基　　金：重庆市教委科学技术研究计划重大项目(KJZD-M202100801).

摘　　要：针对向量自回归模型(VAR)的高维估计问题,结合贝叶斯理论提出了一种融合正态-逆Wishart共轭先验分布的估计方法。在该估计方法中,所提出的模型引入Metropolis-Hastings(MH)算法,从以往数据集中确定先验分布超参数,并通过设定与模型尺寸相关的收缩系数从而进行估计。与传统VAR模型相比,基于贝叶斯理论的估计方法可在保留相关样本信息的同时控制过度拟合,具有较好的稳健性和有效性。此外,在改进的BVAR模型基础上,结合区制转移技术与误差修正模型提出了MS-BVECM模型,该模型能够有效分析经济周期内各变量之间长期与短期均衡状态变化,当短期内经济变量受到波动而与长期均衡状态发生偏离时,误差修正模型机制会使其逐渐重新回到长期均衡状态,以保证模型的稳健性。最后,以重庆市为例,利用所提模型对其能源消费、产业结构升级和经济增长的动态关系进行了分析与预测并提供了可行建议。An estimation method incorporating a normal-inverse Wishart conjugate prior distribution was proposed in conjunction with Bayesian theory to address the problem of high-dimensional estimation of traditional vector autoregressive models(VARs).In this estimation method,the proposed model introduced the Metropolis-Hastings(MH)algorithm to determine the hyperparameters of the prior distribution from the previous dataset and to estimate them by setting the shrinkage coefficients associated with the model dimensions.Compared with traditional VAR models,the estimation method based on Bayesian theory can control overfitting while retaining relevant sample information and has better robustness and effectiveness.In addition,based on the improved BVAR model,the MS-BVECM model was proposed by combining the zone system transfer technique with the error correction model,which can effectively analyze the long-run and short-run equilibrium state changes between the variables during the economic cycle.When the economic variables are subject to fluctuations in the short run and deviate from the long-run equilibrium,the error correction model mechanism will gradually bring them back to the long-run equilibrium to ensure the robustness of the model.Finally,taking Chongqing as an example,the proposed model was used to analyze and predict the dynamic relationship between energy consumption,industrial structure upgrading,and economic growth in Chongqing,and to provide feasible suggestions.

关 键 词：向量自回归模型 正态-逆Wishart共轭先验分布 贝叶斯理论 误差修正模型 能源消费 

分 类 号：O212.7[理学—概率论与数理统计]