Fancy Autocorrelation plot of rugarch package? On page 26 in the rugarch package you can see plots of autocorrelations. I want to use this autocorrelation plots, but not for my final model checking, but for my model identification. So these are NOT the results of a fitted model, but just data (financial returns). So how can I use this autocorrelation plots for my data without fitting a model before? I don't want to use other autocorrelation function, because later I will fit a model and therefore I will have the same plots for all autocorrelation plots. Otherwise I would have differen plot designs for the same idea. How can I do this?
 A: The function responsible for plotting the ACF in the package is called .plot.garchfit.4. It is not hard to modify it to produce plots for time series. 
acf_rugarch = function(x, ...)
{   
    T = length(x)
    insample = 1:T
    xseries = x
    lag.max = as.integer(10*log10(T))
    acfx    = acf(xseries, lag.max = lag.max, plot = FALSE)
    clim0   = qnorm((1 + 0.95)/2)/sqrt(acfx$n.used)
 ylim  = range(c(-clim0, clim0, as.numeric(acfx$acf)[-1]))
    clx     = vector(mode = "character", length = lag.max)
    clx[which(as.numeric(acfx$acf)[-1]>=0)] = "steelblue"
 clx[which(as.numeric(acfx$acf)[-1]<0)] = "orange"
    barplot(height = as.numeric(acfx$acf)[-1], names.arg = as.numeric(acfx$lag)[-1], ylim = 1.2*ylim, col = clx,
            ylab = "ACF", xlab="lag", main = "ACF of Observations", cex.main = 0.8)
    abline(h = c(clim0, -clim0), col = "tomato1", lty = 2)
    abline(h = 0, col = "black", lty = 1)
    box()   
    grid()
}

This function is basically a wrapper for base R function acf. Here is the example:
set.seed(11)

#Generate simple MA(1) process
z<-rnorm(1000)
y<-z+0.5*c(NA,z[-1000])
acf_rugarch(na.omit(y))


A: I found a solution, not perfect but:
One can use no mean equation for the arma model and give this as a input for the standard garch. Then one use the rugarch plot to get the acf of the observation
