Vector Autoregression Forecasting Models: Recent Developments Applied to the U.S. Hog Market
提出一种结合并扩展变量排除技术的向量自回归预测方法，用美国生猪市场季度数据评估1977-1984年样本外预测表现，结果优于其他变量排除法和贝叶斯方法。
Abstract Bayesian estimation and the exclusion of variables are two basic approaches to the improvement of vector autoregression forecasting models. This study presents a method which combines and extends several techniques within the exclusion‐of‐variables approach. Several quarterly hog market models are estimated and out‐of‐sample forecasts from 1977 through 1984 are evaluated. The results suggest the proposed method compares favorably to other exclusion‐of‐variables techniques as well as to the more sophisticated bayesian approaches.