I have a timeseries data where the response variable is autocorrelated in time, like this:

library(tseries) 

testdat <- data.frame(time = as.factor(rep(c(1:10),20)),
                      y = rep(arima.sim(model = list(ar = .75), n = 20, innov = runif(100, 0, 1)),10),
                      group = rep(1:20,10))

Now I would like to test for AR(2), instead of the ar1() argument. Is there a quick fix in either the “time” column in my data or the code?

I have tried going into glmmTMB:::formatVC but could only get the corr value for AR(2). Also there was little luck trying to find the covariance matrix in the results:

m <- glmmTMB(y ~ ar1(time + 0|group) , data = testdat, family = poisson, doFit = F)

View(m)

Am I looking for the right place? Or can I modify the source code to generate AR(2)?