贝叶斯自回归分布滞后模型在经济数据分析中的应用

Application of Bayesian Autoregressive Distribution Lag Model in Economic Data Analysis

在自回归分布滞后模型的基础上,指定合适的先验分布,结合贝叶斯模型思想,建立了贝叶斯自回归分布滞后模型,并借助切片Gibbs抽样方法得到了各个待估参数的后验样本,完成了贝叶斯推断。再以云南省生产总值及第三产业增加值数据为样本进行实证分析。结果显示,用贝叶斯自回归分布滞后模型拟合云南省生产总值数据比传统的ADL模型表现出更好的优良性;云南省第三产业增加值对当年的生产总值产生了较强的支撑作用,占据了生产总值中较大的份额。根据模型分析结果提出云南省未来经济增长点必须向以研发服务和营销服务为重点的新型服务业(第三产业)转移才能维持云南省经济持续稳定增长的产业结构调整的建议。

Firstly,based on the autoregressive distributed lag model,the appropriate prior distributions are designated and the Bayesian autoregressive distributed lag model is established by combining it with the Bayesian modeling theory.Slice Gibbs sampling is used to get samples of posterior waiting estimable parameters and accomplish Bayesian inference.Secondly,the GDP and the added value of the tertiary industry of Yunnan Province are used as samples to accomplish empirical study.The results of Bayesian autoregressive distributed lag model show that it has excellent fitting effect compared with the traditional ADL model and the added value of the tertiary industry contributed greatly to the GDP of Yunnan Province as it accounted for a considerable portion of the latter.Lastly,suggestions regarding adjustment of industrial structure of Yunnan Province are made according to results of analysis of the model.