## a general framework for updating belief functions

Posted in Books, Statistics, University life with tags , , , , , , , , , on July 15, 2013 by xi'an

Pier Giovanni Bissiri, Chris Holmes and Stephen Walker have recently arXived the paper related to Sephen’s talk in London for Bayes 250. When I heard the talk (of which some slides are included below), my interest was aroused by the facts that (a) the approach they investigated could start from a statistics, rather than from a full model, with obvious implications for ABC, & (b) the starting point could be the dual to the prior x likelihood pair, namely the loss function. I thus read the paper with this in mind. (And rather quickly, which may mean I skipped important aspects. For instance, I did not get into Section 4 to any depth. Disclaimer: I wasn’t nor is a referee for this paper!)

The core idea is to stick to a Bayesian (hardcore?) line when missing the full model, i.e. the likelihood of the data, but wishing to infer about a well-defined parameter like the median of the observations. This parameter is model-free in that some degree of prior information is available in the form of a prior distribution. (This is thus the dual of frequentist inference: instead of a likelihood w/o a prior, they have a prior w/o a likelihood!) The approach in the paper is to define a “posterior” by using a functional type of loss function that balances fidelity to prior and fidelity to data. The prior part (of the loss) ends up with a Kullback-Leibler loss, while the data part (of the loss) is an expected loss wrt to l(THETASoEUR,x), ending up with the definition of a “posterior” that is

$\exp\{ -l(\theta,x)\} \pi(\theta)$

the loss thus playing the role of the log-likelihood.

I like very much the problematic developed in the paper, as I think it is connected with the real world and the complex modelling issues we face nowadays. I also like the insistence on coherence like the updating principle when switching former posterior for new prior (a point sorely missed in this book!) The distinction between M-closed M-open, and M-free scenarios is worth mentioning, if only as an entry to the Bayesian processing of pseudo-likelihood and proxy models. I am however not entirely convinced by the solution presented therein, in that it involves a rather large degree of arbitrariness. In other words, while I agree on using the loss function as a pivot for defining the pseudo-posterior, I am reluctant to put the same faith in the loss as in the log-likelihood (maybe a frequentist atavistic gene somewhere…) In particular, I think some of the choices are either hard or impossible to make and remain unprincipled (despite a call to the LP on page 7).  I also consider the M-open case as remaining unsolved as finding a convergent assessment about the pseudo-true parameter brings little information about the real parameter and the lack of fit of the superimposed model. Given my great expectations, I ended up being disappointed by the M-free case: there is no optimal choice for the substitute to the loss function that sounds very much like a pseudo-likelihood (or log thereof). (I thought the talk was more conclusive about this, I presumably missed a slide there!) Another great expectation was to read about the proper scaling of the loss function (since L and wL are difficult to separate, except for monetary losses). The authors propose a “correct” scaling based on balancing both faithfulness for a single observation, but this is not a completely tight argument (dependence on parametrisation and prior, notion of a single observation, &tc.)

The illustration section contains two examples, one of which is a full-size or at least challenging  genetic data analysis. The loss function is based on a logistic  pseudo-likelihood and it provides results where the Bayes factor is in agreement with a likelihood ratio test using Cox’ proportional hazard model. The issue about keeping the baseline function as unkown reminded me of the Robbins-Wasserman paradox Jamie discussed in Varanasi. The second example offers a nice feature of putting uncertainties onto box-plots, although I cannot trust very much the 95%  of the credibles sets. (And I do not understand why a unique loss would come to be associated with the median parameter, see p.25.)

Watch out: Tomorrow’s post contains a reply from the authors!

## a partial review of BISP8 [guest post]

Posted in Statistics, Travel, University life with tags , , , , , , , on June 17, 2013 by xi'an

Chris Drovandi (QUT) sent me his impression on BISP8 that just took place in Milano, Italia (BISP stands for Bayesian inference in stochastic processes):

Here is a review of some of the talks at BISP8. For the other talks I do not have sufficient background to give the talks the justice that they deserve. It was a very enjoyable small workshop with many talks in my areas of interest.

In the first session Vanja Dukic presented bayesian inference of SEIR epidemic DE models and state space models of google flu trends data. In the case of the state space models a particle learning algorithm was developed. The author considered both fixed and random effects for the data in each US state. In the second session, Murali Haran presented a likelihood-free approach for inferring the parameters of a spatio-temporal epidemic model. The speaker used a Gaussian process emulator of the model based on model simulations from a regulator grid of parameter values. The emulator approach is suggested to be less intensive in terms of the number of model simulations compared with abc but is only suitable for low dimensional inference problems (even less so than abc).

In the first session of day 2 Ana Palacios combined the gompertz model with Markov processes to create flexible and realistic stochastic growth models. The resulting model has a difficult likelihood and inference was performed by completing the likelihood creating simple Gibbs moves and by ABC.

There were 3 talks in a row on inference for SDEs. The first, by Simon Särkkä, avoids evaluating an intractable transition density by proposing from another diffusion model and computing importance weights using the girsanov theorem. Next, Samuel Kou used a population MCMC type approach where each chain had a different Euler discretisation. This helps improve mixing for the chain with the finest grid. Moves between chains are complicated by the different dimension for each chain. The author used a filling approach to overcome this. A very interesting aspect of the talk was using information from all chains to extrapolate various posterior quantiles to delta_t is 0 (no discretisation implying the correct posterior). I assume the extrapolation may not work as well for the extreme quantiles. The third talk, by Andrew Golightly, proposed an auxiliary approach to improve PMCMC for these models. This talk was the most technical (for me) so need more time to digest. Following my talk (based on some work here.  And some current work.) was an applied talk using smc2 methodology.

On the final day Alexandros Beskos investigated the use of SMC for Bayesian inference for a high dimensional (static) parameter. SMC is advocated here due to the ease of adaptation relative to MCMC when there is no structure in the model. The base of the approach I believe was that of Chopin (2002).

## Bayes 250 in London

Posted in Books, pictures, Statistics, Travel, University life with tags , , , , , on March 20, 2013 by xi'an

The two-day Bayes 250 Conference at the Royal Statistical Society is now officially announced with the complete programme on the RSS website. With the registration form available as well. A mix of eighteen junior and senior speakers covering the thematic and geographical spectra of UK Bayesian statistics. (It would be difficult not to acknowledge the top position of the United Kingdom in the list of contributions to Bayesian statistics!) Plus an interview of Dennis Lindley (pictured above in one of the rare pictures of Dennis available on the Web) by Tony O’Hagan! Thanks to Chris Holmes for organising this exciting meeting celebrating the 1763 publication of the Essay (with me “tagging along” as a co-organiser).

Here is a blurb I wrote as a presentation (pardon my French!):

2013 marks the 250th anniversary of the publication in Dec. 1763 of “An Essay towards solving a Problem in the Doctrine of Chances” in the Philosophical Transactions of the Royal Society of London, based on notes by Thomas Bayes and edited by Richard Price who submitted the Essay posthumously to Bayes’ death.

This publication is acknowledged as the birth certificate of what is now called Bayesian statistics and the Royal Statistical Society decided to celebrate this important milestone in the story of statistics (and not only UK statistics) by organising a conference on Bayesian statistics. The conference will take place at the RSS Headquarters in Errol Street and will run from June 19, late morning, to June 20, early afternoon. Everyone interested is welcome to present one’s work during the poster session on the afternoon of June 19.

The Royal Statistical Society is looking forward your participation in this event and hopes you will enjoy the variety in the presentations of the programme.

## who’s afraid of the big B wolf?

Posted in Books, Statistics, University life with tags , , , , , , , , , , on March 13, 2013 by xi'an

Aris Spanos just published a paper entitled “Who should be afraid of the Jeffreys-Lindley paradox?” in the journal Philosophy of Science. This piece is a continuation of the debate about frequentist versus llikelihoodist versus Bayesian (should it be Bayesianist?! or Laplacist?!) testing approaches, exposed in Mayo and Spanos’ Error and Inference, and discussed in several posts of the ‘Og. I started reading the paper in conjunction with a paper I am currently writing for a special volume in  honour of Dennis Lindley, paper that I will discuss later on the ‘Og…

“…the postdata severity evaluation (…) addresses the key problem with Fisherian p-values in the sense that the severity evaluation provides the “magnitude” of the warranted discrepancy from the null by taking into account the generic capacity of the test (that includes n) in question as it relates to the observed data”(p.88)

First, the antagonistic style of the paper is reminding me of Spanos’ previous works in that it relies on repeated value judgements (such as “Bayesian charge”, “blatant misinterpretation”, “Bayesian allegations that have undermined the credibility of frequentist statistics”, “both approaches are far from immune to fallacious interpretations”, “only crude rules of thumbs”, &tc.) and rhetorical sleights of hand. (See, e.g., “In contrast, the severity account ensures learning from data by employing trustworthy evidence (…), the reliability of evidence being calibrated in terms of the relevant error probabilities” [my stress].) Connectedly, Spanos often resorts to an unusual [at least for statisticians] vocabulary that amounts to newspeak. Here are some illustrations: “summoning the generic capacity of the test”, ‘substantively significant”, “custom tailoring the generic capacity of the test”, “the fallacy of acceptance”, “the relevance of the generic capacity of the particular test”, yes the term “generic capacity” is occurring there with a truly high frequency. Continue reading

## the anti-Bayesian moment and its passing commented

Posted in Books, Statistics, University life with tags , , , , on March 12, 2013 by xi'an

Here is a comment on our rejoinder “the anti-Bayesian moment and its passing” with Andrew Gelman from Deborah Mayo, comment that could not make it through as a comment:

You assume that I am interested in long-term average properties of procedures, even though I have so often argued that they are at most necessary (as consequences of good procedures), but scarcely sufficient for a severity assessment. The error statistical account I have developed is a statistical philosophy. It is not one to be found in Neyman and Pearson, jointly or separately, except in occasional glimpses here and there (unfortunately). It is certainly not about well-defined accept-reject rules. If N-P had only been clearer, and Fisher better behaved, we would not have had decades of wrangling. However, I have argued, the error statistical philosophy explicates, and directs the interpretation of, frequentist sampling theory methods in scientific, as opposed to behavioural, contexts. It is not a complete philosophy…but I think Gelmanian Bayesians could find in it a source of “standard setting”.

You say “the prior is both a probabilistic object, standard from this perspective, and a subjective construct, translating qualitative personal assessments into a probability distribution. The extension of this dual nature to the so-called “conventional” priors (a very good semantic finding!) is to set a reference … against which to test the impact of one’s prior choices and the variability of the resulting inference. …they simply set a standard against which to gauge our answers.”

I think there are standards for even an approximate meaning of “standard-setting” in science, and I still do not see how an object whose meaning and rationale may fluctuate wildly, even in a given example, can serve as a standard or reference. For what?

Perhaps the idea is that one can gauge how different priors change the posteriors, because, after all, the likelihood is well-defined. That is why the prior and not the likelihood is the camel. But it isn’t obvious why I should want the camel. (camel/gnat references in the paper and response).