Introduction To Astrostatistics

• Activity Number: 387
• Type: Invited
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Eric Feigelson*+
• Companies: Penn State University
• Keywords: astrostatistics ; image ; Bayesian ; truncation ; time series ; spatial processes

Abstract

After a very brief review of the historical interplay between statistics and astronomy, this talk will discuss the state of astrostatistics today. Astronomers confront a bewildering variety of methodological challenges in analyzing and interpreting their data. Data can be in images, spectra, time series, and multivariate datasets. The samples can have either homogeneous or heterogeneous populations. Some datasets are sparse, while others reach petabyte scales. Astronomers need diverse methodologies including nonparametrics, regression, multivariate analysis and classification, time series analysis, spatial point processes, and image processing. Some particular challenges are common in astronomy: heteroscedastic measurement errors, truncation and censoring, and regression using complex models from astrophysical theory. Bayesian approaches to modeling are becoming popular. Astronomers are poorly trained in statistics, but interactions are improving with astrostatistical conferences, research groups, international organizations, and summer schools. Astrostatistics is thus emerging as an important cross-disciplinary enterprise engaging researchers from both communities.