search

grd349 / PBjam

A repo for our peak baggin code and tips on jam
MIT License
17 stars 6 forks source link

kde.plot_spectrum() xlim is inconsistent #214

Closed nielsenmb closed 4 years ago

nielsenmb commented 4 years ago

It should be focused around numax, but appears to sometimes just take the full frequency range.

nielsenmb commented 4 years ago

This is being fixed in a PR I'm working on.

I relation to this, need some opinions on the corner plot of the stage result in relation to prior: At start: star_corner

After KDE: star_kde

After Asy_peakbag: star_asy_fit

grd349 commented 4 years ago

Hi Martin,

Could you explain a little more. For the data, why is the numax so bimodal? How are you generating the black data - is it all the points from the prior or is it the output of the KDE (if it is the output of the KDE then something is not working)?

nielsenmb commented 4 years ago

Black is the prior, red is the latest stage values for the relevant parameters, so if it's called from the star class it plots the input values on top of the prior, if it's called from the KDE class it plots the KDE results etc. (doesn't work for peakbag of course). 

This star has numax at 312muHz (it's Eps Ret, a TESS target), putting it just beyond the Kepler long cadence Nyquist.