[R-SIG-Finance] rugarch singular issue

[alexios galanos]

Hi Jia,

Not sure that the use of "inv" is the right solution since A IS probably 
singular or close to such. This means that something went wrong in the
optimization. Would you mind sharing a reproducible example? I can try
and investigate a little more.

Thanks,

Alexios

On 2/16/20 3:56 PM, Jia Shao wrote:
> Hi Alexios,
> 
> I am working on ‘rugarch’ package recently and sometimes have
> Error in robustvcv(fun = f, pars = ipars[estidx, 1], nlag = nlag, hess = fit$hessian, : object 'B' not found
> 
> Error in solve.default(A) :
>    system is computationally singular: reciprocal condition number = 1.91379e-16
> 
> which is related to the file
> rugarch-numderiv.R line 140
> 
> robustvcv = function(fun, pars, nlag = 0, hess, n, ...)
> 
> ...
> 
>          Ainv = try( solve(A), silent = TRUE )
> 
> Because 'solve(A)’ think the matrix A is singular. Instead of using solve(A), using inv(A) available in the package ‘matlib’ will solve the issue.
> 
> Do you think this is a reasonable change in the package?
> 
> Best regards
> Jia Shao
> 
> Top 15 UK University – five years running
> Guardian University Guide 2016-2020
> 
> 
> 1st for Overseas Student Experiences
> based on student trips abroad from HESA 2017/18 UK data
> 
> 
> 2nd for Teaching Excellence
> Times Higher Education UK metrics ranking 2017
> 
> 
> University of the Year for Student Experience 2019
> The Times and Sunday Times Good University Guide 2019
> 
> NOTICE
> 
> This message and any files transmitted with it is intended for the addressee only and may contain information that is confidential or privileged. Unauthorised use is strictly prohibited. If you are not the addressee, you should not read, copy, disclose or otherwise use this message, except for the purpose of delivery to the addressee.
> 
> Any views or opinions expressed within this e-mail are those of the author and do not necessarily represent those of Coventry University.
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-SIG-Finance using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions should go.
>