在检验GARCH（1,2）模型的时候，想用广义混合检验。    也就是R程序中的gBox（）函数，检验标准残差。
    R帮助如下：

Generalized Portmanteau Tests for GARCH ModelsDescriptionPerform a goodness-of-fit test for the GARCH model by checking whether the standardized residuals are iid based on the ACF of the absolute residuals or squared residuals.
UsagegBox(model, lags = 1:20, x, method = c("squared", "absolute")[1], plot = TRUE)Arguments

model	fitted model from the garch function of the tseries library
lags	a vector of maximum ACF lags to be used in the test
x	time series data to which the GARCH model is fitted
method	"squared": test is based on squared residuals; "absolute": test is based on absolute residuals
plot	logical variable, if TRUE, the p-values of the tests are plotted

Value

lags	lags in the input
pvalue	a vector of p-values of the tests
method	method used
x	x

Author(s)Kung-Sik Chan
References"Time Series Analysis, with Applications in R" by J.D. Cryer and K.S. Chan
Exampleslibrary(tseries)data(CREF)r.cref=diff(log(CREF))*100m1=garch(x=r.cref,order=c(1,1))summary(m1)#gBox(m1,x=r.cref,method='squared')############################################

我的这段程序如下gBox(g09,x=DEC09_R1,method='squared')但是运行过后老是出错如：“错误于filter(M, filter = beta, method = "recursive", sides = 1, init = rep(sigma2,：dims [product 2]与对象长度[1]不匹配”查看residuals(g09)残差序列时，发现前几个值是NA，于是我用例子里面的garch（1,1）模型又做了一遍，还是这样，这个和空值有关么？？？为什么会有这样的错误提示啊？？求解！！！
如果不用广义混合检验，只用ACF行不行丫？总感觉不完善……
还有模型拟合以后，在之后20多阶的时候出现一个相关性是不是能说明有异常值的存在？