統計数学セミナー
Seminar on Probability and Statistics |
Home : Archive [ 2003 to 04 ] [ 2004 to 05 ] [ 2005 to 06 ] [ 2006 to 07 ] [ 2007 to 08 ] [ 2008 to 09 ] [ 2009 to 10 ] [ 2010 to 11 ] [ 2011 to 12 ] [ 2012 to 13 ] [ 2013 to 14 ] [ 2014 to 15 ] |
Previous Seminar : Next Seminar |
Seminar on Probability and Statistics Thursday October 18 2012 Tokyo 006 3:15-4:25 pm
Quasi-Bayesian analysis of nonparametric instrumental variables models
加藤 賢悟 / KATO, Kengo 広島大学大学院理学研究科数学専攻 / Department of Mathematics, Graduate School of Science, Hiroshima University Abstract This paper aims at developing a quasi-Bayesian analysis
of the nonparametric instrumental variables model, with a focus on the
asymptotic properties of quasi-posterior distributions. In this paper,
instead of assuming a distributional assumption on the data generating
process, we consider a quasi-likelihood induced from the conditional
moment restriction, and put priors on the function-valued parameter.
We call the resulting posterior quasi-posterior, which corresponds to
``Gibbs posterior'' in the literature. Here we shall focus on sieve
priors, which are priors that concentrate on finite dimensional sieve
spaces. The dimension of the sieve space should increase as the sample
size. We derive rates of contraction and a non-parametric Bernstein-von
Mises type result for the quasi-posterior distribution, and rates of
convergence for the quasi-Bayes estimator defined by the posterior
expectation. We show that, with priors suitably chosen, the
quasi-posterior distribution (the quasi-Bayes estimator) attains the
minimax optimal rate of contraction (convergence, respectively). These
results greatly sharpen the previous related work.
|
Previous Seminar : Next Seminar | Seminar on Probability and Statistics |