Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 416 - Open Problems in Astrostatistics
Type: Topic Contributed
Date/Time: Wednesday, August 10, 2022 : 10:30 AM to 12:20 PM
Sponsor: Section on Physical and Engineering Sciences
Abstract #323217
Title: Topics for Statistical Advances for Use in Astronomy
Author(s): Herman Marshall*
Companies: MIT
Keywords: calibration; shrinkage; Bayesian

I will discuss some areas where statistical advances have been made in astronomical data analysis. The first area -- dubbed Concordance -- has been a subject of collaboration between astronomers and statisticians for seven years. The first Concordance paper was published in a statistics journal, the second was in a journal devoted to astronomy research, indicating the nature of the cross-disciplinary effort. Concordance is a framework for assessing systemat errors in the effective areas of telescopes that we have particularly applied to a group of X-ray instruments. The method was developed because X-ray telescopes are launched into orbit where they are no longer available for calibration using ground-based standards. The key statistical technique used in this project is multiplicative shrinkage with the first-ever use of priors on instrument effective areas combined in a Bayesian analysis. The project is the subject of further development that will be discussed. In X-ray polarimetry, one development involves the statistical handling of neural networks trained on simulated event tracks as the first step in data analysis.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program