Archive for blogging

prior sensitivity of the marginal likelihood

Posted in Books, pictures, Statistics, University life with tags , , , , , , , , , , , , , on June 27, 2022 by xi'an

Fernando Llorente and (Madrilene) coauthors have just arXived a paper on the safe use of prior densities for Bayesian model selection. Rather than blaming the Bayes factor, or excommunicating some improper priors, they consider in this survey solutions to design “objective” priors in model selection. (Writing this post made me realised I had forgotten to arXive a recent piece I wrote on the topic, based on short courses and blog pieces, for an incoming handbook on Bayesian advance(ment)s! Soon to be corrected.)

While intrinsically interested in the topic and hence with the study, I somewhat disagree with the perspective adopted by the authors. They for instance stick to the notion that a flat prior over the parameter space is appropriate as “the maximal expression of a non-informative prior” (despite depending on the parameterisation). Over bounded sets at least, while advocating priors “with great scale parameter” otherwise. They also refer to Jeffreys (1939) priors, by which they mean estimation priors rather than testing priors. As uncovered by Susie Bayarri and Gonzalo Garcia-Donato. Considering asymptotic consistency, they state that “in the asymptotic regime, Bayesian model selection is more sensitive to the sample size D than to the prior specifications”, which I find both imprecise and confusing,  as my feeling is that the prior specification remains overly influential as the sample size increases. (In my view, consistency is a minimalist requirement, rather than “comforting”.) The argument therein that a flat prior is informative for model choice stems from the fact that the marginal likelihood goes to zero as the support of the prior goes to infinity, which may have been an earlier argument of Jeffreys’ (1939), but does not carry much weight as the property is shared by many other priors (as remarked later). Somehow, the penalisation aspect of the marginal is not exploited more deeply in the paper. In the “objective” Bayes section, they adhere to the (convenient but weakly supported) choice of a common prior on the nuisance parameters (shared by different models). Their main argument is to develop (heretic!) “data-based priors”, from Aitkin (1991, not cited) double use of the data (or setting the likelihood to the power two), all the way to the intrinsic and fractional Bayes factors of Tony O’Hagan (1995), Jim Berger and Luis Pericchi (1996), and to the expected posterior priors of Pérez and Berger (2002) on which I worked with Juan Cano and Diego Salmeròn. (While the presentation is made against a flat prior, nothing prevents the use of another reference, improper, prior.) A short section also mentions the X-validation approach(es) of Aki Vehtari and co-authors.

[de]quarantined by slideshare

Posted in Books, pictures, Statistics, University life with tags , , , , , , , , , , , , , , , , on January 11, 2021 by xi'an

A follow-up episode to the SlideShare m’a tuer [sic] saga: After the 20 November closure of my xianblog account and my request for an explanation, I was told by Linkedin that a complaint has been made about one of my talks for violation of copyright. Most surprisingly, at least at first, it was about the slides for the graduate lectures I gave ten years ago at CREST on (re)reading Jaynes’ Probability Theory. While the slides contain a lot of short quotes from the Logic of Science, somewhat necessarily since I discuss the said book, there are also many quotes from Jeffreys’ Theory of Probability and “t’is but a scratch” on the contents of this lengthy book… Plus, the pdf file appears to be accessible on several sites, including one with an INRIA domain. Since I had to fill a “Counter-Notice of Copyright Infringement” to unlock the rest of the depository, I just hope no legal action is going to be taken about this lecture. But I remain puzzled at the reasoning behind the complaint, unwilling to blame radical Jaynesians for it! As an aside, here are the registered 736 views of the slides for the past year:

quarantined by slideshare

Posted in Books, pictures, Statistics, University life with tags , , , , , , , on November 26, 2020 by xi'an

Just found out that SlideShare has closed my account for “violating SlideShare Terms of Service”! As I have no further detail (and my xianblog account is inaccessible) I have contacted SlideShare to get an explanation for this inexplicable cancellation and hopefully (??) can argue my case. If not the hundred plus slide presentations that were posted there and linked on the ‘Og will become unavailable. I seem to remember this happened to me once before so maybe there is hope to invert the decision presumably made by an AI using the wrong prior or algorithm!

xi’an’s number [repost]

Posted in Statistics with tags , , , , on August 18, 2019 by xi'an

“extremely damaging and slanderous blog article”

Posted in Books, pictures, University life with tags , , , , , on April 13, 2019 by xi'an

Yesterday was a first for the ‘Og in that my university legal department received a complaint from a company about one of the posts, that among other things was considered as an “offending” and “extremely damaging and slanderous commentary” on the services proposed by this company, including posting the unsolicited marketing email it has sent me earlier and which induced this comment of mine.  Written in superb legalese of course. As my point had been made when the blog was posted, a while ago, and as I saw no point in bothering my legal department representatives or wasting further time on such nonsense, I removed this terribly damaging entry and hope the poor dears have recovered by now… It’s “a mere attempt at bloggin’, nothing more”, right?!

%d bloggers like this: