eigen() for symmetric matrices (ADMM_algorithm.R)

[KrystynaGrzesiak]

Is the assumption about the symmetry of C_eigen right?

[mstaniak]

@mmaku any thoughts?

[mmaku]

sorry, I forgot to answer.
Yeah, I am sure that it must be symetric under some assumptions, which are always true in case of precision matrix, I remember I checked that.
I'll give you theoretical proof, but give me some (little) time.

[mstaniak]

I see @mmaku - the problem is that @KrystynaGrzesiak found cases where it wasn't symmetric. Maybe it's a good starting point for debugging. @KrystynaGrzesiak can you provide a reproducible example of calculations failing (or being too slow?) due to non-symmetric matrix?

[mmaku]

We tried to set up am meeting, but unfortunately it did not work out. We will try to arrange a meeting and discuss the problem.