R help - Feb 2015 - Rolling window VAR (vector autoregressive) estimation

Hi,

I am completely new to R. I have a dataset of several thousand, weekly
observations with ten variables. I would like to create a program (algorithm)
that does a rolling window VAR regression of window width "w" (for
example, a window width of 200 weeks) for all of the variables within the model.
The best package I can find to execute VAR regressions seems to be the
"vars" package
(http://cran.r-project.org/web/packages/vars/index.html). For my algorithm, I
would like to retreive certain objects from the class "varest"
generated by the "vars" package after a call to "VAR."  More
specifically, within each rolling window I'm interested in retreiving a
matrix of the variance-covariance residuals and a matrix of the estimated
coefficients for the lagged endogenous variables.

There are several examples within this discussion board about estimating rolling
window regressions with linear regression models, but I have not been able to
find any algorithms for rolling window regressions with a more complicated
regression like the vector autoregressive model.

As I have several thousand observations, I would appreciate advise on creating
efficient algorithms to execute the rolling window regression for this VAR
execution.

Thanks,

Wesley