Open
Description
It's not necessary to compute it if you don't call
rand
orlogpdf
shrugOne approach that doesn't touch the structure of
FiniteGP
would be to add a methodMvNormal(::FiniteGP)
that returns the correspondingMvNormal
distribution which could then be used instead ofFiniteGP
if you don't want to recompute the cholesky decomposition, e.g., if you callrand
orlogpdf
multiple times.
Ooh, I like that. Maybe we have a @require
in MeasureTheory that adds the MvNormal
method. Then we can use the faster representation too, should work out great!
Originally posted by @cscherrer in JuliaGaussianProcesses/AbstractGPs.jl#232 (comment)
What do you think @mschauer?
Metadata
Metadata
Assignees
Labels
No labels