I worked on putting references into my similarity-of-objects document (how do you determine that two different objects are identical in their measurable properties>?), and tweaking the words, with the hope that I will have something postable to the arXiv soon.
Hi David; just stumbled upon this old entry of yours. I wonder if you're aware of this:
ReplyDeleteProbabilistic Record Linkage in Astronomy: Directional Cross-Identification and Beyond
http://www.annualreviews.org/doi/abs/10.1146/annurev-statistics-010814-020231?journalCode=statistics
It's about Bayesian coincidence assessment/cross-matching/cross-identification/whatever-you-want-to-call-it. The origins are in work I did with Ira Wasserman back in the mid 90s on assessing evidence for repeating of GRBs. We came up with a quite general framework for Bayesian coincidence assessment (distinguishing spatio-temporal coincidences from genuine associations), briefly described in some proceedings papers. After numerous proposals to take the idea further (met with some of the most obnoxious reviewer reports I've gotten), we finally gave up on it in the mid 00s. Then, just a few years later, Tamas Budavari and Alex Szalay independently rediscovered the main idea. It's now the cross-matching algorithm used for SDSS. The review article tells some of the story, including computational tricks, other astro applications, and relating it to work elsewhere on record linkage when there is feature uncertainty.