Welcome to Brews! This is where I store results of analyses (typically, posterior samples, marginal likelihood estimates, etc) that you might like to use. You are welcome to use them for any purpose you like, as long as you acknowledge where they came from. However, if you use them for research that you plan to publish, please contact me - I might want to contribute writing (e.g. about the methodology).


In 2009, Dennis Stello and I wrote this paper on fitting the quasi-periodic oscillations seen in asteroseismology with a model that has a gaussian process for the conditional prior for the data.

Recently, I implemented a very similar idea but within the nicer framework of DNest4 and with Dan Foreman-Mackey's celerite package to speed up the likelihood evaluations. Here is the output of a run on the same xi Hydrae data from the 2009 paper: xi_hydrae.tar.gz. Try running to generate some plots.