University of Birmingham > Talks@bham > Astrophysics Talks Series > The Unified Statistical Analysis of Populations of Sources: Advantages of “Shrinkage Estimates” in Astronomy

The Unified Statistical Analysis of Populations of Sources: Advantages of “Shrinkage Estimates” in Astronomy

Add to your list(s) Download to your calendar using vCal

  • UserDavid van Dyk (Imperial)
  • ClockWednesday 29 April 2015, 14:30-15:30
  • HousePhysics West 117.

If you have a question about this talk, please contact Ilya Mandel.

Astronomical studies often involve samples or populations of sources. The parameters describing the sources can either be fit to each source in a separate analysis, or all be fit in a single unified analysis. The latter strategy allows us to incorporate the population distribution into a coherent statistical model and exhibits distinct statistical advantages. In particular, objects with smaller error bars and well-constrained parameters allow us to estimate the population distribution, which in turn can be used to better estimate the weakly-constrained parameters associated with objects with larger error bars. The fitted values of such weakly-constrained parameters will “shrink towards’’ the population mean, and are thus called “shrinkage estimates’’. This talk describes both frequentist and Bayesian advantages of shrinkage estimates and illustrates how they can be used in astronomy. In the first of two examples we estimate the absolute magnitudes of a SDSS sample of 288 Type Ia Supernovae using shrinkage estimates and illustrate how they differ from naive estimates. In the second example, we use photometric magnitudes of a sample of galactic halo white dwarfs to simultaneously obtain shrinkage estimates of the stellar ages and an estimate the age of the halo.

This talk is part of the Astrophysics Talks Series series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


Talks@bham, University of Birmingham. Contact Us | Help and Documentation | Privacy and Publicity.
talks@bham is based on from the University of Cambridge.