This is designed as a means to convey Statistics concepts to astronomers in an easy fashion. Of the many comments we got from the participants at the AstroStat Workshop at HEAD 2004 in New Orleans, a major theme was that it was necessary to have a place where Astronomers could consult with Statisticians, and also a place which could act as a repository of editorial commentary, akin to Bad Astronomy, except that it would deal with bad Astrostatistics.
The CHASC Astrostatistics Collaboration was initiated in 1997 in response to challenges posed at the Statistical Challenges in Modern Astronomy II meeting at Penn State (Siemiginowska et al. 1997), and is now hosted by the CfA, and the Statistics departments of Harvard and UC (Irvine). It provides a forum for discussion of outstanding statistical problems in Astrophysics. Goals include developing papers on the interface of statistics and astronomy and incorporating state-of-the-art statistical methods into CIAO. Many regular members of CHASC have expertise in Bayesian statistical methods in general and computational methods such as MCMC samplers and the EM algorithm in particular. However the group’s perspective is more toward practical likelihood-based hybrid solutions to real problems rather than toward philosophical purity. CHASC has received funding through Chandra, NSF and the NASA AISR program.
Members of CHASC have included, at one time or another, Paul Baines, James Chiang, Alanna Connors, Paul Edlefson, David Esch, Peter Freeman, Rima Izem, Hosung Kang, Vinay Kashyap, Tae Yeon Kwon, Hyunsook Lee, Alan Lenarcic, Jingchen Liu, Xiao-li Meng, Taeyoung Park, Rostislav Protassov, Aneta Siemiginowska, Jeff Scargle, Epaminondas Sourlas, David van Dyk, Yue Wu, C. Alex Young, Yaming Yu, and Andreas Zezas.
In addition to interesting research topics in astrostatistics, current regular postings from the SLOG are based on quotes, arXiv, and MADS. You’ll find enchanting quotes from statisticians and astronomers, reflecting their rich insights on the subject matter toward enhanced science. Also, you’ll find the most up to date publications related to astrostatistics that is retrieved from arXiv (http://arxiv.org).
What is arXiv? Here’s a link to Wikipedia: arXiv. Recently, statistics community joined arXiv [see IMS Bulletin Vol. 36(4), 2007, p.3 and Jim Pitman's IMS Presidential Address, Open Access to Professional Information].
MADS is an acronym, Missing in ADS (a popular site for astronomers covering various databases in astronomy, astronomers’ google scholar, or JSTOR but having a way longer history and better. See http://ads.harvard.edu). While reading astro-ph, I realize that so many charming statistics, statistical methodology, algorithms, and notions are almost never used in astronomy, whereas they are fully utilized in other subjects, from agriculture to zoology. How to utilize and apply these statistics are left for astronomers but I hope these introductions could help astronomers be inspired when they were stuck with simple linear regression, chi-square minimization, optimization, feature detection, brute force Monte Carlo, interpreting p-values, distinguishing between model selection and hypothesis testing, disentangling between model fitting and model estimation, handling massive data from machine learning perspective, not from human eye inspection, and so on.
SLOG is an acronym of Science Weblog.
Contact at chasc.astrostat at gmail dot com