Linear optimum prediction filter for MRW (multifractal random walk) model

I have computed estimates in R for the MRW parameters (lamda, R, sigma). I am now at the point where I have to compute volatility forecasts for this model.

I know it is done via the optimum linear prediction filter. Which in turn uses the R functions acf() and the acf2ar(). The acf2ar() function

returns a matrix. I am not sure how to interpret or analysis the input of this matrix and how to proceed from here.

Any suggestions would be highly appreciated!

Thanks