Statistics – Applications
Scientific paper
Jan 2009
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2009aas...21321104l&link_type=abstract
American Astronomical Society, AAS Meeting #213, #211.04; Bulletin of the American Astronomical Society, Vol. 41, p.283
Statistics
Applications
Scientific paper
Source properties reported in astronomical survey catalogs often have significant uncertainties, arising directly via measurement error, or indirectly when source properties of interest must be inferred via indicators with significant "population scatter" (e.g., luminosity indicators, photometric redshifts). The interplay of source uncertainties and survey selection effects underlies the familiar Malmquist and Lutz-Kelker biases. But failure to account for source uncertainties can corrupt population-level inferences even when selection effects are negligible or absent. I will use simple examples to elucidate the nature of such "scatter biases" and motivate a general framework for handling them, known in the statistics literature as hierarchical or multilevel modeling. I will sketch how Bayesian multilevel modeling provides a powerful, unified perspective on population modeling, spanning a range of applications from the simple (e.g., estimating 1-D densities, such as flux distributions, or redshift distributions along a pencil beam) to the complex (e.g., leading to adaptive generalizations of Malmquist and Lutz-Kelker corrections).
No associations
LandOfFree
Handling Source Uncertainties in Surveys via Multilevel Modeling does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Handling Source Uncertainties in Surveys via Multilevel Modeling, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Handling Source Uncertainties in Surveys via Multilevel Modeling will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1696536