Astrostatistics

Members of the Statistics Section have a strong interest in astrostatistics, the application and development of statistical methods to data analysis problems in astronomy and astrophysics.  This activity is facilitated in part through the links to the Imperial Centre for Inference and Cosmology (ICIC), hosted by the Astrophysics Group in the Department of Physics.

The main contributions from Statistics are to develop principled statistical methods (most often Bayesian) and numerical techniques (e.g., sampling algorithms) to increase both the reliability and precision with which astronomical data are interpreted. This approach has led to a greater understanding of the white dwarf population (van Dyk; Mortlock), efficient outlier search methods (Gandy; Mortlock), more accurate results from X-ray spectra (van Dyk), the reliable classification of astronomical sources into stars and galaxies (Mortlock; Gandy) and the discovery of the most distant known quasars (Mortlock).

But astrostatistics is not just an area in which statisticians can make contributions to a research field; it is also extremely useful as a way of getting access to real data-sets (generally with well characterised, but heteroskedastic noise properties) on which to test methods and algorithms. Please contact any members of this group if you wish to explore the opportunities for doing this.