Estimatng error in ARIMA(p,d,q)

2 views (last 30 days)
MINA
MINA on 12 Aug 2014
Edited: MINA on 12 Aug 2014
Hello,
I am trying to model my data with ARIMA(1,2,12) and since the variance is not stationary I have also included GARCH (2,3). so I have saved all the parameter of my model in a variable called mdl. I have used 1000 time steps as initial values and I want to do a 10 time step ahead for forecasting and then I want to compute the residuals. this is my code:
[res,V,LOGL] = infer(mdl,data(1:1000));
[YF, YMSE,VV] = forecast(mdl,10,'Y0',data(1:1000),'E0',res,'V0',V);
ERROR=data(1001:1010)-YF;
My question is if I have to double difference my data to compute the ERROR or should I don't need to do that. Because D=2 is my ARIMA so I don't know what forecasts gives me, the data after being differentiated two times or the without differencing. Thanks

Answers (0)

Categories

Find more on Conditional Mean Models in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!