How can i manage the garchfit problem?

Hi all, I'm running iterations in which I generate returns with garchsim and then i estimate parameters with garchfit.This process works for about 80,000 iterations but suddenly i receive following massage:

??? Error using ==> svd Input to SVD must not contain NaN or Inf.

Error in ==> pinv at 29 [U,S,V] = svd(A,0);

Error in ==> qpsub at 461 projSD = pinv(projH)*(-Zgf);

Error in ==> nlconst at 694 [SD,lambda,exitflagqp,outputqp,howqp,ACTIND] ...

Error in ==> fmincon at 758 [X,FVAL,LAMBDA,EXITFLAG,OUTPUT,GRAD,HESSIAN]=...

Error in ==> garchfit at 904 [coefficients,logL,exitFlag,output,lambda] = ...

I wonder that garchfit has this problem in some cases and i didn't receive this error massage in past when i worked with different data. So, my question is how i can solve this problem?

Any suggestion on how to correct this problem would be greatly appreciated. Thanks a lot to all in advance.

M.Gorji

[Shashank Prasanna]

Can you share what reproduction code you have? In particular the model you are generating your data with.

[Gorji]

Dear Shashank Here's part of my code:

ret=xlsread('S&P.csv','I2:I3501');
q=1;
nboot=1000;
Is=10000;
Spec=garchset('r',1,'p',1,'q',1,'VarianceModel','EGARCH','Display','off');

for i=1:2000

Ret=ret(i:Is+(i-1));

[Coeff,~,~,Innovations,Sigmas]= garchfit(Spec,Ret);
Sr= Innovations./ Sigmas;
while q<=nboot  
Sr_b=Sr(unidrnd(Is,2*Is,1));                                          %Sr:Standardized Residuals & Sd:Standard deviation
[~,~,Ret_b] = garchsim(Coeff,2*Is,1,Sr_b,[],[]...
,Innovations,Sigmas,Ret);

%Estimating ARMA-EGARCH model parameters for simulated time series(Ret_b)
Coeff_b= garchfit(Spec,Ret_b(Is+1:end));
q=q+1;
end
q=1;
end

I face the problem when i want to estimate parameters based on Ret_b (simulated data) with garchfit in some cases. I hope you can help me. Thanks a lot in advance