Forecasting Vector Autoregressions with Bayesian Priors
论证了在预测向量自回归模型时使用贝叶斯先验的合理性，并发现非对称随机游走先验在预测美国生猪市场五个序列时优于其他三种时间序列模型。
Abstract The paper explores the justification for, and application of, Bayesian priors in forecasting a vector autoregression. A nonsymmetric, random‐walk prior outperforms three alternative time‐series representations in forecasting five series of the U.S. hog market.