LevenbergMarquardt problem in Multivariate GARCH

I tried to use the code, proposed by Kevin Sheppard in his UCSD GARCH Toolbox. There is a part from the file "scalar_bekk_mvgarch": "[t k]=size(data); garchmat=zeros(k,1+p+q); options=optimset('fmincon'); options=optimset(options,'TolCon',1e-3,'Display','off','Diagnostics','off','LargeScale','off', 'LevenbergMarquardt','on');"

So Matlab gives an error message: "The LevenbergMarquardt option is no longer valid." What should I change there without any harm to sense of the code?

Answer 1

[Matt J]

I tend to think that you should just get rid of

options=optimset(options,'TolCon',1e3,'Display','off','Diagnostics','off','LargeScale','off', 'LevenbergMarquardt','on');

and see how well the default settings do. TolCon=1e-3 looks very strange to me

[Erdni]

thank you, I just got rid of "'LevenbergMarquardt','on'". And everything run pretty good.