Hennawi and I spent part of the day pair-coding a simple toy code to infer the spectroscopic noise model given a few (extracted, one-dimensional) spectra of the same object taken with the same spectrograph. Hennawi wants a precise noise model for the SDSS spectrographs for Lyman-alpha forest clustering and IGM studies. We figured out a quasi-correct Gibbs sampling method, that draws posterior samples of the object's true spectrum, and given each of those, subsequently samples in noise-model parameter space. It seems to work! Now to run on real data, which contain a few additional complications (like non-trivial interpolation and non-Gaussian noise components like cosmic rays).
Why only quasi-correct? Usually it's possible tweak things to make an actually-correct method. Usually by making some random choices and/or rejecting some moves.
ReplyDeleteIf you manage to get the bluest wavelengths right, you get a cookie.
ReplyDeleteIain: Great comment: I realized right after you wrote this that I can correct my inexact sampling with a small rejection adjustment. Nice!
ReplyDelete