Replies: 1 comment
-
Hi @EvanClaes. I don't think it should matter as long as you're using the same scaling factors for the observations and the observation noise (within an iteration). We often just use the |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi,
I'm wondering what is the best way to specify observation noise when using a FixedNoiseGP together with NEI. The case I am working on has a fixed amount of observation noise calculated as a percentage from the true (noiseless) observation. Thus for each observation I know the exact amount of noise.
However for fitting the FixedNoiseGP the observations are standardized. The value with which they are scaled varies from iteration to iteration as more queries are evaluated, and thus the observation noiselevels (which are also scaled) vary as well.
Is this harmful for NEI? Would I be better off choosing a fixed scaling value for the standardization so that the specified noise levels for FixedNoiseGP are fixed as well?
Beta Was this translation helpful? Give feedback.
All reactions