Nonparametric Bayesian aggregation for massive data

Document Type

Article

Publication Date

9-1-2019

Abstract

We develop a set of scalable Bayesian inference procedures for a general class of nonparametric regression models. Specifically, nonparametric Bayesian inferences are separately performed on each subset randomly split from a massive dataset, and then the obtained local results are aggregated into global counterparts. This aggregation step is explicit without involving any additional computation cost. By a careful partition, we show that our aggregated inference results obtain an oracle rule in the sense that they are equivalent to those obtained directly from the entire data (which are computationally prohibitive). For example, an aggregated credible ball achieves desirable credibility level and also frequentist coverage while possessing the same radius as the oracle ball.

Identifier

85077511250 (Scopus)

Publication Title

Journal of Machine Learning Research

e-ISSN

15337928

ISSN

15324435

Volume

20

Grant

DMS-1712907

Fund Ref

Office of Naval Research

This document is currently not available here.

Share

COinS