Closed Prof-ThiagoOliveira closed 2 years ago
Thanks for working on this. I think I'm going to go in a different way however. I think I'm going to implement this check in the rotMat
function along with the test for symmetry. I'll also document this function so that I can point the user to an explanation for the warning.
I suspect that three cases to be the most common for triggering this warning. The first is just a mistake in the input. The second would be intended when the user is setting a correlation to 1. The last would be the trickiest for a user where what they are seeing is likely the bi-product of the estimation process. For this case, I was thinking of suggesting the user investigate matrix bending. Perhaps by citing this paper. Any thoughts?
great! =D
I suggest you add a warning message when correlation matrices are not positive semi-definite (as explained in the opened issue - https://github.com/gaynorr/AlphaSimR/issues/43#issue-1146083980).