[英]Fitting a DCC Garch Model in R
I'm trying to run a DCC Multivariate GARCH Model. 我正在尝试运行DCC多变量GARCH模型。 When I run the model, it shows only the statistics of the GARCH part, but i need the statistics of the VAR part too.
当我运行模型时,它只显示GARCH部分的统计数据,但我也需要VAR部分的统计数据。 Does anyone know how to do it?
有谁知道怎么做?
This is my routine: 这是我的惯例:
library(parallel)
library(rugarch)
library(rmgarch)
library("tseries")
library("zoo")
library("forecast")
library("FinTS")
library("vars")
library("MTS"
uspec.n = multispec(replicate(3, ugarchspec(mean.model = list(armaOrder = c(1,1)))))
dcc.11mn = dccspec(uspec.n, VAR = TRUE, lag = 4, lag.max = 12, dccOrder = c(1, 1), distribution = 'mvnorm')
fit.2.11mn = dccfit(dcc.11mn, data = B [1:192, 1:3])
And this is my result: 这是我的结果:
---------------------------------*
* DCC GARCH Fit *
*---------------------------------*
Distribution : mvnorm
Model : DCC(1,1)
No. Parameters : 26
[VAR GARCH DCC UncQ] : [12+9+2+3]
No. Series : 3
No. Obs. : 192
Log-Likelihood : -777.3919
Av.Log-Likelihood : -4.05
Optimal Parameters
-----------------------------------
Estimate Std. Error t value Pr(>|t|)
[rtc].omega 0.000103 0.000092 1.128480 0.259117
[rtc].alpha1 0.299707 0.172126 1.741210 0.081647
[rtc].beta1 0.594874 0.212706 2.796699 0.005163
[ei].omega 0.014031 0.006434 2.180585 0.029214
[ei].alpha1 0.793922 0.223400 3.553810 0.000380
[ei].beta1 0.205078 0.111009 1.847400 0.064689
[i2].omega 682.273357 833.174186 0.818884 0.412852
[i2].alpha1 0.218414 0.061933 3.526637 0.000421
[i2].beta1 0.726249 0.082908 8.759650 0.000000
[Joint]dcca1 0.000000 0.000000 0.000032 0.999974
[Joint]dccb1 0.923723 0.121097 7.627947 0.000000
But what about the statistics of the VAR? 但是VAR的统计数据呢?
See the varfit
function of the rmgarch package, for example for lags=4
. 请参阅rmgarch包的
varfit
函数,例如,对于lags=4
。
V<-varxfit(data, 4, constant = TRUE)
show(V)
and you must correct the dccspec function as below:
dcc.11mn = dccspec(uspec.n, VAR = TRUE, lag = 4, lag.max = 12, dccOrder = c(1, 1), distribution = 'mvnorm', VAR.fit=V, out.sample=4)
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.