Continue running for loop for values that do not given an error

Suppose I have the following line of code

Mdl_vol(1) = egarch('GARCHLags',1,'ARCHLags',1);

Mdl_vol(2) = egarch('GARCHLags',1:2,'ARCHLags',1);

Mdl_vol(3) = egarch('GARCHLags',1:3,'ARCHLags',1);

Mdl_vol(4) = egarch('GARCHLags',1,'ARCHLags',1:2);

Mdl_vol(5) = egarch('GARCHLags',1:2,'ARCHLags',1:2);

Mdl_vol(6) = egarch('GARCHLags',1:3,'ARCHLags',1:2);

Mdl_vol(7) = egarch('GARCHLags',1,'ARCHLags',1:3);

Mdl_vol(8) = egarch('GARCHLags',1:2,'ARCHLags',1:3);

Mdl_vol(9) = egarch('GARCHLags',1:3,'ARCHLags',1:3);

% AIC i stocks j models

for i = 1:20

for j = 1:9

EstMdl = estimate(Mdl_vol(j),monthlyerrorsi(:,i),'Display','off');

results = summarize(EstMdl);

AIC(j,i) = results.AIC;

end

end

Now I get the following error

Estimated EGARCH model is invalid.

Caused by:

Error using egarch/validateModel (line 789)

GARCH polynomial is not stationary.

So for some i = 1:20 any of the 9 models is invalid. However, I am wondering if there is a way to have the once that give an error display a 0 or NaN value. And for the models that do not given an error to obtain the right values.

Answer 1

[Voss]

monthlyerrorsi = randn(100,20); % random monthlyerrorsi

Mdl_vol(1)     = egarch('GARCHLags',1,'ARCHLags',1);
Mdl_vol(2)     = egarch('GARCHLags',1:2,'ARCHLags',1);
Mdl_vol(3)     = egarch('GARCHLags',1:3,'ARCHLags',1);
Mdl_vol(4)     = egarch('GARCHLags',1,'ARCHLags',1:2);
Mdl_vol(5)     = egarch('GARCHLags',1:2,'ARCHLags',1:2);
Mdl_vol(6)     = egarch('GARCHLags',1:3,'ARCHLags',1:2);
Mdl_vol(7)     = egarch('GARCHLags',1,'ARCHLags',1:3);
Mdl_vol(8)     = egarch('GARCHLags',1:2,'ARCHLags',1:3);
Mdl_vol(9)     = egarch('GARCHLags',1:3,'ARCHLags',1:3);
% AIC i stocks j models
for i = 1:20
    for j = 1:9
        try
            EstMdl = estimate(Mdl_vol(j),monthlyerrorsi(:,i),'Display','off');
            results = summarize(EstMdl);
            AIC(j,i) = results.AIC;
        catch ME
            disp(sprintf('%s -- Occurred at (j,i) = (%d,%d)',ME.message,j,i));
            AIC(j,i) = NaN;
        end
    end
end

Estimated EGARCH model is invalid. -- Occurred at (j,i) = (6,3)
Estimated EGARCH model is invalid. -- Occurred at (j,i) = (6,6)
Estimated EGARCH model is invalid. -- Occurred at (j,i) = (5,17)

AIC

AIC = 9×20

1.0e+20 * 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 NaN 0.0000 0.0000 0.0000 0.0000 0.0000 NaN 0.0000 0.0000 NaN 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

[Wietze Zijpp]

This is exactly what I was looking for !

[Voss]

Perfect! Glad to help