Prev: Siebel CRM On Demand -- 10+ Months Contract
Next: Two Factor Factorial Design
From: Ergin Erdem on 30 Jun 2010 16:08
Hi, I have been working on the wind speed project, and I am trying to
fit the data over 3 months priod GARCH(1,1)-M models where ARMA
Compornent consist of an ARMA (2,3) model. I have 3 month of hourly
average wind speed data, and I import it to SAS by export function. I
have defined variable y as a average hourly wind speed, and run the
following code.

-----------------------------------------------------------------------------------------------------------------------------------------------------------
proc model data = Hourly maxiter=1000;
parms ar1 ar2 ma1 ma2 ma3 mu arch0 arch1 garch1 phi gamma;


h.y = arch0 + arch1*xlag(resid.y**2,mse.y) +
garch1*xlag(h.y,mse.y) +phi*xlag(-resid.y,mse.y);


h.y=h;

y = mu + ar1 * zlag1(y - mu) + ar2 * zlag2(y-mu) + ma1 *
zlag1(resid.y)+ ma2 * zlag2(resid.y) + ma3 * zlag3(resid.y)
+gamma*(sqrt(h));

fit y / method = marquardt fiml ;
run;
quit



------------------------------------------------------------------------------------------------------------------------------------------------------------------

SAS returned the following error message.

NOTE: At OLS Iteration 10 CONVERGE=0.001 Criteria Met.
NOTE: At SUR Iteration 0 CONVERGE=0.001 Criteria Met.
ERROR: At FIML Iteration 0 there are 0 nonmissing observations
available. At least 20 are
needed to compute the next iteration.
NOTE: The data set WORK.FORECAST has 0 observations and 4 variables.
185 quit



Can anybody help me what is wrong, and what should be my research
direction, the GMM estimation also provides similar error message. If
you can help me on that it would be great. Thank you.

Ergin
 | 
Pages: 1
Prev: Siebel CRM On Demand -- 10+ Months Contract
Next: Two Factor Factorial Design