Question

提前致谢。在this post中，我询问了如何在VAR模型中选择特定滞后的问题。快速回复并提供“限制”信息。和＆＃39; coef＆＃39;函数我能够成功运行VAR模型，具有我想要的特定滞后。但是，我需要使用限制VAR模型进行预测的代码是什么？

我的代码示例如下：

 ##Attempt to Restrict VAR Coefficients
 ##VAR has 5 lags with three variables plus constant and 11 seasonal dummies.

 library("vars")
 var1 <- VAR(DVARmat, p = 5, type ="const", season = 12)
 restrict <- matrix (c(1,0,0,1,0,1,0,0,0,0,0,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0,0,
                       1,0,0,1,0,1,0,0,0,0,0,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0,0,
                       1,0,0,1,0,1,0,0,0,0,0,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0,0),
                     nrow = 3, ncol = 27, byrow = T)
 var1_restrict <- coef(restrict(var1, method ="man", resmat = restrict))
 var1_restrict

我知道普通VAR后的预测代码，但似乎无法将限制的VAR推入其中。再次感谢。

Answer 1

生成受限系数矩阵restrict后，您可以predict使用restrict(...)，因为restrict(...)也会返回类varest的对象：

 ##Attempt to Restrict VAR Coefficients
     ##VAR has 5 lags with three variables plus constant and 11 seasonal dummies.

         library("vars")
         var1 <- VAR(DVARmat, p = 5, type ="const", season = 12)
         restrict <- matrix (c(1,0,0,1,0,1,0,0,0,0,0,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0,0,
                               1,0,0,1,0,1,0,0,0,0,0,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0,0,
                               1,0,0,1,0,1,0,0,0,0,0,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0,0),
                             nrow = 3, ncol = 27, byrow = T)
         var1_restrict <- coef(restrict(var1, method ="man", resmat = restrict))
         var1_restrict

         expostrestrict <- predict(restrict(var1, method="man", resmat = restrict), n.ahead = 13, ci=.95)

请注意restrict(var1, method="man", resmat = restrict)是可以生成的对象，因此如果有人愿意，也可以使用以下内容：

restrict_var <- restrict(var1, method="man", resmat = restrict)
expostrestrict <- predict(restrict_var, n.ahead = 13, ci=.95)

具有特定VAR模型滞后的预测

1 个答案: