rndWishart error with scalar covariance matrix

Dear forum,

when I type

rndWishart(1,2,5)

in Gauss 13, I get the error message ``(0) : error G0648 : Incorrect input: Covariance matrix not positive definite'', although the scalar covariance ``matrix'' (second input, = 2) is obviously positive definite.

Is this problem related to the input being a scalar? I get no error message when I type ``2*eye(2)'' instead of ``2'' in the command above. On the other hand, the help file does not seem to exclude the case that the covariance matrix is a scalar.

Thanks in advance,

Fabian

PS: I am using Kernel 13.0.9.

1 Answer

0

The error is indeed being returned because the incoming covariance matrix is a scalar. I will forward this to development and get their thoughts on the matter.

link

Aptech

70