search

MNGuenther / allesfitter

allesfitter is a convenient wrapper around the packages ellc (light curve and RV models), dynesty (static and dynamic nested sampling) emcee (Markov Chain Monte Carlo sampling) and celerite (Gaussian Process models).
MIT License
60 stars 36 forks source link

Fitting offset + gp for RVs #20

Closed jvines closed 3 years ago

jvines commented 3 years ago

Hi,

Is it possible to fit both an offset and a GP to the RVs?

Cheers,

Jose V.

MNGuenther commented 3 years ago

Hi Jose, Yes!

But first I would check -- is it needed? The GP can handle small offsets on its own. But it wouldn't handle the full systemic offset of non-normalized RVs. So you could just remove the systemic RV offset (by subtracting it manually from your data files).

If you fit the GP because you see stellar variability, then use the "stellar_variability" terms, which are not instrument dependent.

If you are really sure you need/want to do the GP and an offset at the same time, I would recommend doing the offsets as "baselines" (per instrument) and the GP as "stellar variability". See the tutorial here: https://github.com/MNGuenther/allesfitter/tree/master/tutorials/06_transits_and_rvs_with_stellar_variability

jvines commented 3 years ago

In that case, yes, I need it! I haven't normalized them at all! How can I fit for the offset and GP??

MNGuenther commented 3 years ago

See above (I just finished editing, GitHub sent my first sentence by accident when I just wanted to hit enter for a new paragraph)

jvines commented 3 years ago

Thank you very much!

Also, real quick, I just noticed that in the ns_derived_latex_table.txt the companion density units appear as \$rho_\mathrm{b}$ (cgs) instead of $\rho_\mathrm{b}$ (cgs)

MNGuenther commented 3 years ago

Cheers, just fixed the typo. Will come with the next version!