2 Stats papers on astro-ph today
There are two statistics papers on astro-ph. Check them out:
1/ http://arxiv.org/abs/0709.0596
Title: Bayesian Inversion of Stokes Profiles
Authors: A. Asensio Ramos (1), M. J. Martinez Gonzalez (2), J. A. Rubino-Martin (1) ((1) IAC, (2) LERMA)
Comments: 15 pages, 12 figures, accepted for publication in A&A
2/ http://arxiv.org/abs/0709.0711
Title: Bayesian posterior classification of planetary nebulae according to the Peimbert types
Authors: C. Quireza (1), H.J. Rocha-Pinto (2), W.J. Maciel (3) ((1) Observatorio Nacional/MCT, (2) Observatorio do Valongo/UFRJ, (3) Instituto de Astronomia, Geofisica e Ciencias Atmosfericas/USP)
Comments: 26 pages, 4 figures, 6 tables. Accepted for publication in Astronomy and Astrophysics.
vlk:
Heh, the one about the Stokes Profiles is practically a repeat of Kashyap & Drake (1998, ApJ, 503, 450). The only difference is in the models, and they seem to use a more complicated prior. I am assuming that Ramos et al. learnt their MCMC from Neal (1993) too, since they don’t seem to know about our paper. Somebody should tell Radford Neal what an effect he is having on astrophysics!
Hyunsook, take a look at their eqns 9 and 10
(btw Aneta, the hyperlink from the second paper [0711) goes to the first one [0596])
09-09-2007, 11:24 amhlee:
I am writing summaries on these papers and a few other Bayesian papers appeared in the past month. After I saw those two equations, I felt amazed and amused (the latter mainly due to my lack of diversity in various methodologies and their applications). I already made comment about it and please correct me if my understanding is wrong once I push the button, publish.
By the way, another arxiv paper in the slog, Bayesian Star Formation has о‡^2 as well. Beyond surprise, I haven’t seriously tried to validate statistical justification on their approaches (connecting Bayesian and о‡^2) since I do not know any statistical studies on both topics.
09-10-2007, 12:06 am