How to reverse a prediction obtained from a seasonal difference of time series in Matlab?

Hi everyone

I make a prediction using a weekly differencing data. the weekly differencing time series data has been obtained using the following lines of codes. the data is half hour data

hence 7*48=336.

D12 = LagOp({1 -1},'Lags',[0,336]);

dY = filter(D12,loadr);

After the data transformation I perform a SARMAX regression using the following lines of codes

[beta, betaci, res2] = regress(dY, Xres);

[beta, betaci, res2] = regress(dY, [Xres,lagmatrix(res2, lags)]); with Xres the seasonal AR and exogenous predictors. While res2 is the MA predictor.

Now I can simulate the model for other time period. My question is how can I transform the result gotten from the model in order to remove the lag operator D12 and get the real value.

please help me.

thanks