Consistency of Bayes estimators of a binary regression function
Coram, Marc ; Lalley, Steven P.
Ann. Statist., Tome 34 (2006) no. 1, p. 1233-1269 / Harvested from Project Euclid
When do nonparametric Bayesian procedures “overfit”? To shed light on this question, we consider a binary regression problem in detail and establish frequentist consistency for a certain class of Bayes procedures based on hierarchical priors, called uniform mixture priors. These are defined as follows: let ν be any probability distribution on the nonnegative integers. To sample a function f from the prior πν, first sample m from ν and then sample f uniformly from the set of step functions from [0,1] into [0,1] that have exactly m jumps (i.e., sample all m jump locations and m+1 function values independently and uniformly). The main result states that if a data-stream is generated according to any fixed, measurable binary-regression function f0≢1/2, then frequentist consistency obtains: that is, for any ν with infinite support, the posterior of πν concentrates on any L1 neighborhood of f0. Solution of an associated large-deviations problem is central to the consistency proof.
Publié le : 2006-06-14
Classification:  Consistency,  Bayes procedure,  binary regression,  large deviations,  subadditivity,  62A15,  62E20
@article{1152540748,
     author = {Coram, Marc and Lalley, Steven P.},
     title = {Consistency of Bayes estimators of a binary regression function},
     journal = {Ann. Statist.},
     volume = {34},
     number = {1},
     year = {2006},
     pages = { 1233-1269},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1152540748}
}
Coram, Marc; Lalley, Steven P. Consistency of Bayes estimators of a binary regression function. Ann. Statist., Tome 34 (2006) no. 1, pp.  1233-1269. http://gdmltest.u-ga.fr/item/1152540748/