摘要
由于经济环境的多变,使得经济预测面临数据量少的建模难题,贝叶斯方法对小样本数据建模问题具有明显优势。本文在共轭条件似然函数"矩阵正态-Wishart分布"意义下,首先讨论了向量自回归模型的贝叶斯分析方法,得到了模型参数的后验分布与一步预测分布。其次,给出了分量方程的对应结果,说明了模型阶数的推断方法。最后,列出了计算步骤,并作为应用,对上海房地产价格指数数据进行预测建模,取得了较好效果。
Economical system is in change frequently, so there is always not enough observations to set up the forecasting model. Bayesian approach is better than traditional for small sample size problem. This paper studies Bayesian approach for vector autoregressions with conjugated conditional likelihood function, matrix normal - Wishart distribution. First, the posterior distributions of model parameters and the one-step-ahead predictive distribution are obtained. Second, the results about relative component equations are given and the inference method for model order is discussed. Third, as application, the forecasting model of Shanghai real estate price index is set up and show well performance.
出处
《数理统计与管理》
CSSCI
北大核心
2010年第6期1060-1066,共7页
Journal of Applied Statistics and Management
基金
上海市教育委员会重点科研项目(编号:06ZZ34)
上海理工大学引进人才科研启动经费资助项目
关键词
贝叶斯方法
时间序列
向量自回归模型
经济预测
房地产价格指数
Bayesian analysis, time series, vector autoregressions, economic forecasting, real estate price index