Logo image
Back
Bayesian Quasar Selection and the Quasar Luminosity Function
Journal article   Peer reviewed

Bayesian Quasar Selection and the Quasar Luminosity Function

Gordon T Richards
Classification and Discovery in Large Astronomical Surveys (AIP Conference Proceedings Volume 1082), v 1082
01 Jan 2008
DOI: https://doi.org/10.1063/1.3059053

Abstract

I summarize our work on the construction of deep photometric quasar catalogs using a non-parametric Bayesian classifier based on kernel density estimation. Using only 5-band optical data from the Sloan Digital Sky Survey (SDSS), we have identified nearly 1,000,000 photometric quasars candidates over 0 < z < 5 in 8400 deg2. Overall the catalog is nearly 80% efficient (quasar candidates:true quasars), increasing to 97% for z < 2.2 UVX objects. Using ~30 deg2 of overlap between SDSS and Spitzer-IRAC we extend our algorithm to 8-D color selection and identify 5500 photometric quasars candidates, including up to 1200 type 2 (obscured) quasars. Our method has the potential for achieving quasar densities as high as 1000 deg-2 over large areas of sky. In addition to selection of quasars, we further determine photometric redshifts for our sample. From the SDSS 5-band photometry, we are able to achieve accuracies as high at Dz+/-0.3 for ~83% of the sample. Including MIR data from IRAC improves this fraction to ~94%. The classification accuracy and quasar density in these catalogs are enabling statistical analyses of quasar clustering and the quasar luminosity function that have hitherto been impossible.

Metrics

4 Record Views

Details

Logo image