Manually forecasting with known parameters
Posted: Tue Apr 05, 2011 11:09 pm
Hi All,
I'm trying to figure out that how can I get y^ predictions by using known GARCH/ARIMA coefficients. For instance I run a model like ;
mean equation
Y= b0 + b1(x1) + b2(x2) + ar(1) + ar(4) + ma(5)
variance equation
e^2=a0+a1(e^2(-1))
Estimation Command:
=========================
ARCH(1,0,BACKCAST=0.7,DERIV=AA) y C x1 x2 AR(1) MA(5) AR(4)
Estimation Equation:
=========================
y = C(1) + C(2)*x1 + C(3)*x2 + [AR(1)=C(4),AR(4)=C(5),MA(5)=C(6),BACKCAST=10,ESTSMPL="10 243"]
GARCH = C(7) + C(8)*RESID(-1)^2
Substituted Coefficients:
=========================
y = -0.00295415901785 - 0.624932092896*x1 - 0.213207549753*x2 + [AR(1)=0.711619941237,AR(4)=0.281671400578,MA(5)=-0.7899083843,BACKCAST=10,ESTSMPL="10 243"]
GARCH = 5.23407278082e-05 + 0.0757712993063*RESID(-1)^2
Since I know all that parameters (b0,b1,b2,ar(1),ar(4),ma(5),a0,a1) how can I get future predictions out of this model manually. What I would like to say is that I want to extract whole model to excel and get manually static and dynamic forecasts from it. I already know how to extract AR process and GARCH/ARIMA theory but although I tried to decode model I couldn't reach Eviews' own forecast results.I think that I stucked on identifying initial values,backkast smoothing,etc and then everything was bricked. Do anybody have any of example about making manual forecasts with known GARCH/ARMA parameters using excel. I said excel but that's why I know programming with several languages, for sure it can be eviews or any program code as well.
Thanks in advance.
I'm trying to figure out that how can I get y^ predictions by using known GARCH/ARIMA coefficients. For instance I run a model like ;
mean equation
Y= b0 + b1(x1) + b2(x2) + ar(1) + ar(4) + ma(5)
variance equation
e^2=a0+a1(e^2(-1))
Estimation Command:
=========================
ARCH(1,0,BACKCAST=0.7,DERIV=AA) y C x1 x2 AR(1) MA(5) AR(4)
Estimation Equation:
=========================
y = C(1) + C(2)*x1 + C(3)*x2 + [AR(1)=C(4),AR(4)=C(5),MA(5)=C(6),BACKCAST=10,ESTSMPL="10 243"]
GARCH = C(7) + C(8)*RESID(-1)^2
Substituted Coefficients:
=========================
y = -0.00295415901785 - 0.624932092896*x1 - 0.213207549753*x2 + [AR(1)=0.711619941237,AR(4)=0.281671400578,MA(5)=-0.7899083843,BACKCAST=10,ESTSMPL="10 243"]
GARCH = 5.23407278082e-05 + 0.0757712993063*RESID(-1)^2
Since I know all that parameters (b0,b1,b2,ar(1),ar(4),ma(5),a0,a1) how can I get future predictions out of this model manually. What I would like to say is that I want to extract whole model to excel and get manually static and dynamic forecasts from it. I already know how to extract AR process and GARCH/ARIMA theory but although I tried to decode model I couldn't reach Eviews' own forecast results.I think that I stucked on identifying initial values,backkast smoothing,etc and then everything was bricked. Do anybody have any of example about making manual forecasts with known GARCH/ARMA parameters using excel. I said excel but that's why I know programming with several languages, for sure it can be eviews or any program code as well.
Thanks in advance.