I am starting to get a knowledge about the Bayesian methods in high dimensional settings. The following two references are what I am getting at.

Hjort, Nils Lid. "Bayesian approaches to non-and semiparametric density estimation." Preprint series. Statistical Research Report http://urn. nb. no/URN: NBN: no-23420 (1994).

Müller, Peter, and Fernando A. Quintana. "Nonparametric Bayesian data analysis." Statistical science (2004): 95-110.

Interestingly enough, most Bayesian nonparametrics/parametrics reference did not mention high dimensional setting in full details. The only work that I know of is Chap.9 of :

Ghosh, Jayanta K., Mohan Delampady, and Tapas Samanta. An introduction to Bayesian analysis: theory and methods. Springer Science & Business Media, 2007.

This is a nice short chapter with emphasis in classical hypothesis testing. However, its material is obviously a bit out-dated given the research progress in this field. A book-length treatment assuming a relatively high-level of pre-knowledge is

Frigessi, Arnoldo, et al., eds. Statistical Analysis for High-Dimensional Data: The Abel Symposium 2014. Vol. 11. Springer, 2016.

And most high-dimensional related introduction did not give enough time to Bayesian modelling in high dimensional setting.

My question is: Are there any reference/introductory paper that allows readers to grasp a panoramic view of Bayesian procedures in high dimensional settings? Preferably it should be readable and gives a big picture instead of falling into too much technicalities.



1 Answer 1


Because it has been nine months, I want to start by answering, "No." But I want to hear about it if there's a great survey paper or book!

Then I want to point out two papers that introduce methods for parametric Bayesian modeling in high dimensions.

First, Expectation Propagation can help by using a distributed system for doing parametric Bayesian inference on "big data", including high dimensionality ("hundreds of dimensions").

"Expectation Propagation as a Way of Life". Gelman et al., 2017 http://www.stat.columbia.edu/~gelman/research/unpublished/ep_arxiv.pdf

Also, Stochastic Variational Inference can handle high dimensional data.

"Stochastic Variational Inference". Hoffman, Blei, Wang, and Paisley, 2013 https://arxiv.org/pdf/1206.7051.pdf

  • 1
    $\begingroup$ thanks for your input! I really hope to see one since I am working in the field for quite a while and still have not seen reasonable amount of it. Will some of Rubin's recent work servethis purpose? $\endgroup$
    – Henry.L
    Jun 2, 2017 at 22:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.