## Bayesian parameter estimation versus model comparison

Posted in Books, pictures, Statistics with tags , , , , , , on December 5, 2016 by xi'an

John Kruschke [of puppies’ fame!] wrote a paper in Perspectives in Psychological Science a few years ago on the comparison between two Bayesian approaches to null hypotheses. Of which I became aware through a X validated question that seemed to confuse Bayesian parameter estimation with Bayesian hypothesis testing.

“Regardless of the decision rule, however, the primary attraction of using parameter estimation to assess null values is that the an explicit posterior distribution reveals the relative credibility of all the parameter values.” (p.302)

After reading this paper, I realised that Kruschke meant something completely different, namely that a Bayesian approach to null hypothesis testing could operate from the posterior on the corresponding parameter, rather than to engage into formal Bayesian model comparison (null versus the rest of the World). The notion is to check whether or not the null value stands within the 95% [why 95?] HPD region [modulo a buffer zone], which offers the pluses of avoiding a Dirac mass at the null value and a long-term impact of the prior tails on the decision, with the minus of replacing the null with a tolerance region around the null and calibrating the rejection level. This opposition is thus a Bayesian counterpart of running tests on point null hypotheses either by Neyman-Pearson procedures or by confidence intervals. Note that in problems with nuisance parameters this solution requires a determination of the 95% HPD region associated with the marginal on the parameter of interest, which may prove a challenge.

“…the measure provides a natural penalty for vague priors that allow a broad range of parameter values, because a vague prior dilutes credibility across a broad range of parameter values, and therefore the weighted average is also attenuated.” (p. 306)

While I agree with most of the critical assessment of Bayesian model comparison, including Kruschke’s version of Occam’s razor [and Lindley’s paradox] above, I do not understand how Bayesian model comparison fails to return a full posterior on both the model indices [for model comparison] and the model parameters [for estimation]. To state that it does not because the Bayes factor only depends on marginal likelihoods (p.307) sounds unfair if only because most numerical techniques to approximate the Bayes factors rely on preliminary simulations of the posterior. The point that the Bayes factor strongly depends on the modelling of the alternative model is well-taken, albeit the selection of the null in the “estimation” approach does depend as well on this alternative modelling. Which is an issue if one ends up accepting the null value and running a Bayesian analysis based on this null value.

“The two Bayesian approaches to assessing null values can be unified in a single hierarchical model.” (p.308)

Incidentally, the paper briefly considers a unified modelling that can be interpreted as a mixture across both models, but this mixture representation completely differs from ours [where we also advocate estimation to replace testing] since the mixture is at the likelihood x prior level, as in O’Neill and Kypriaos.

## JSM 2015 [day #4]

Posted in pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , on August 13, 2015 by xi'an

My first session today was Markov Chain Monte Carlo for Contemporary Statistical Applications with a heap of interesting directions in MCMC research! Now, without any possible bias (!), I would definitely nominate Murray Pollock (incidentally from Warwick) as the winner for best slides, funniest presentation, and most enjoyable accent! More seriously, the scalable Langevin algorithm he developed with Paul Fearnhead, Adam Johansen, and Gareth Roberts, is quite impressive in avoiding computing costly likelihoods. With of course caveats on which targets it applies to. Murali Haran showed a new proposal to handle high dimension random effect models by a projection trick that reduces the dimension. Natesh Pillai introduced us (or at least me!) to a spectral clustering that allowed for an automated partition of the target space, itself the starting point to his parallel MCMC algorithm. Quite exciting, even though I do not perceive partitions as an ideal solution to this problem. The final talk in the session was Galin Jones’ presentation of consistency results and conditions for multivariate quantities which is a surprisingly unexplored domain. MCMC is still alive and running!

The second MCMC session of the morning, Monte Carlo Methods Facing New Challenges in Statistics and Science, was equally diverse, with Lynn Kuo’s talk on the HAWK approach, where we discovered that harmonic mean estimators are still in use, e.g., in MrBayes software employed in phylogenetic inference. The proposal to replace this awful estimator that should never be seen again (!) was rather closely related to an earlier solution of us for marginal likelihood approximation, based there on a partition of the whole space rather than an HPD region in our case… Then, Michael Betancourt brilliantly acted as a proxy for Andrew to present the STAN language, with a flashy trailer he most recently designed. Featuring Andrew as the sole actor. And with great arguments for using it, including the potential to run expectation propagation (as a way of life). In fine, Faming Liang proposed a bootstrap subsampling version of the Metropolis-Hastings algorithm, where the likelihood acknowledging the resulting bias in the limiting distribution.

My first afternoon session was another entry on Statistical Phylogenetics, somewhat continued from yesterday’s session. Making me realised I had not seen a single talk on ABC for the entire meeting! The issues discussed in the session were linked with aligning sequences and comparing  many trees. Again in settings where likelihoods can be computed more or less explicitly. Without any expertise in the matter, I wondered at a construction that would turn all trees, like  into realizations of a continuous model. For instance by growing one branch at a time while removing the MRCA root… And maybe using a particle like method to grow trees. As an aside, Vladimir Minin told me yesterday night about genetic mutations that could switch on and off phenotypes repeatedly across generations… For instance  the ability to glow in the dark for species of deep sea fish.

When stating that I did not see a single talk about ABC, I omitted Steve Fienberg’s Fisher Lecture R.A. Fisher and the Statistical ABCs, keeping the morceau de choix for the end! Even though of course Steve did not mention the algorithm! A was for asymptotics, or ancilarity, B for Bayesian (or biducial??), C for causation (or cuffiency???)… Among other germs, I appreciated that Steve mentioned my great-grand father Darmois in connection with exponential families! And the connection with Jon Wellner’s LeCam Lecture from a few days ago. And reminding us that Savage was a Fisher lecturer himself. And that Fisher introduced fiducial distributions quite early. And for defending the Bayesian perspective. Steve also set some challenges like asymptotics for networks, Bayesian model assessment (I liked the notion of stepping out of the model), and randomization when experimenting with networks. And for big data issues. And for personalized medicine, building on his cancer treatment. No trace of the ABC algorithm, obviously, but a wonderful Fisher’s lecture, also most obviously!! Bravo, Steve, keep thriving!!!

## Bayesian model averaging in astrophysics [guest post]

Posted in Books, pictures, Statistics, University life with tags , , , , , , , , , , , , , on August 12, 2015 by xi'an

.[Following my posting of a misfiled 2013 blog, Ewan Cameron told me of the impact of this paper in starting his own blog and I asked him for a guest post, resulting in this analysis, much deeper than mine. No warning necessary this time!]

Back in February 2013 when “Bayesian Model Averaging in Astrophysics: A Review” by Parkinson & Liddle (hereafter PL13) first appeared on the arXiv I was a keen, young(ish) postdoc eager to get stuck into debates about anything and everything ‘astro-statistical’. And with its seemingly glaring flaws, PL13 was more grist to the mill. However, despite my best efforts on various forums I couldn’t get a decent fight started over the right way to do model averaging (BMA) in astronomy, so out of sheer frustration two months later I made my own soapbox to shout from at Another Astrostatistics Blog. Having seen PL13 reviewed recently here on Xian’s Og it feels like the right time to revisit the subject and reflect on where BMA in astronomy is today.

As pointed out to me back in 2013 by Tom Loredo, the act of Bayesian model averaging has been around much longer than its name; indeed an early astronomical example appears in Gregory & Loredo (1992) in which the posterior mean representation of an unknown signal is constructed for an astronomical “light-curve”, averaging over a set of constant and periodic candidate models. Nevertheless the wider popularisation of model averaging in astronomy has only recently taken place through a variety of applications in cosmology: e.g. Liddle, Mukherjee, Parkinson & Wang (2006) and Vardanyan, Trotta & Silk (2011).

In contrast to earlier studies like Gregory & Loredo (1992)—or the classic review on BMA by Hoeting et al. (1999)—in which the target of model averaging is typically either a utility function, a set of future observations, or a latent parameter of the observational process (e.g. the unknown “light-curve” shape) shared naturally by all competing models, the proposal of cosmological BMA studies is to produce a model-averaged version of the posterior for a given ‘shared’ parameter: a so-called “model-averaged PDF”. This proposal didn’t sit well with me back in 2013, and it still doesn’t sit well with me today. Philosophically: without a model a parameter has no meaning, so why should we want to seek meaning in the marginalised distribution of a parameter over an entire set of models? And, practically: to put it another way, without knowing the model ‘label’ to which a given mass of model-averaged parameter probability belongs there’s nothing much useful we can do with this ‘PDF’: nothing much we can say about the data we’ve just analysed and nothing much we can say about future experiments. Whereas the space of the observed data is shared automatically by all competing models, it seems to me to be somehow “un-Bayesian” to place the further restriction that the parameters of separate models share the same scale and topology. I say “un-Bayesian” since this mode of model averaging suggests a formulation of the parameter space + prior pairing stronger than the statement of one’s prior beliefs for the distribution of observable data given the model. But I would be happy to hear arguments from the other side in the comments box below … ! Continue reading

## a unified treatment of predictive model comparison

Posted in Books, Statistics, University life with tags , , , , , , , , , on June 16, 2015 by xi'an

“Applying various approximation strategies to the relative predictive performance derived from predictive distributions in frequentist and Bayesian inference yields many of the model comparison techniques ubiquitous in practice, from predictive log loss cross validation to the Bayesian evidence and Bayesian information criteria.”

Michael Betancourt (Warwick) just arXived a paper formalising predictive model comparison in an almost Bourbakian sense! Meaning that he adopts therein a very general representation of the issue, with minimal assumptions on the data generating process (excluding a specific metric and obviously the choice of a testing statistic). He opts for an M-open perspective, meaning that this generating process stands outside the hypothetical statistical model or, in Lindley’s terms, a small world. Within this paradigm, the only way to assess the fit of a model seems to be through the predictive performances of that model. Using for instance an f-divergence like the Kullback-Leibler divergence, based on the true generated process as the reference. I think this however puts a restriction on the choice of small worlds as the probability measure on that small world has to be absolutely continuous wrt the true data generating process for the distance to be finite. While there are arguments in favour of absolutely continuous small worlds, this assumes a knowledge about the true process that we simply cannot gather. Ignoring this difficulty, a relative Kullback-Leibler divergence can be defined in terms of an almost arbitrary reference measure. But as it still relies on the true measure, its evaluation proceeds via cross-validation “tricks” like jackknife and bootstrap. However, on the Bayesian side, using the prior predictive links the Kullback-Leibler divergence with the marginal likelihood. And Michael argues further that the posterior predictive can be seen as the unifying tool behind information criteria like DIC and WAIC (widely applicable information criterion). Which does not convince me towards the utility of those criteria as model selection tools, as there is too much freedom in the way approximations are used and a potential for using the data several times.

## ABC by population annealing

Posted in Statistics, University life with tags , , , , , , , , on January 6, 2015 by xi'an

The paper “Bayesian Parameter Inference and Model Selection by Population Annealing in System Biology” by Yohei Murakami got published in PLoS One last August but I only became aware of it when ResearchGate pointed it out to me [by mentioning one of our ABC papers was quoted there].

“We are recommended to try a number of annealing schedules to check the influence of the schedules on the simulated data (…) As a whole, the simulations with the posterior parameter ensemble could, not only reproduce the data used for parameter inference, but also capture and predict the data which was not used for parameter inference.”

Population annealing is a notion introduced by Y Iba, the very same IBA who introduced the notion of population Monte Carlo that we studied in subsequent papers. It reproduces the setting found in many particle filter papers of a sequence of (annealed or rather tempered) targets ranging from an easy (i.e., almost flat) target to the genuine target, and of an update of a particle set by MCMC moves and reweighing. I actually have trouble perceiving the difference with other sequential Monte Carlo schemes as those exposed in Del Moral, Doucet and Jasra (2006, Series B). And the same is true of the ABC extension covered in this paper. (Where the annealed intermediate targets correspond to larger tolerances.) This sounds like a traditional ABC-SMC algorithm. Without the adaptive scheme on the tolerance ε found e.g. in Del Moral et al., since the sequence is set in advance. [However, the discussion about the implementation includes the above quote that suggests a vague form of cross-validated tolerance construction]. The approximation of the marginal likelihood also sounds standard, the marginal being approximated by the proportion of accepted pseudo-samples. Or more exactly by the sum of the SMC weights at the end of the annealing simulation. This actually raises several questions: (a) this estimator is always between 0 and 1, while the marginal likelihood is not restricted [but this is due to a missing 1/ε in the likelihood estimate that cancels from both numerator and denominator]; (b) seeing the kernel as a non-parametric estimate of the likelihood led me to wonder why different ε could not be used in different models, in that the pseudo-data used for each model under comparison differs. If we were in a genuine non-parametric setting the bandwidth would be derived from the pseudo-data.

“Thus, Bayesian model selection by population annealing is valid.”

The discussion about the use of ABC population annealing somewhat misses the point of using ABC, which is to approximate the genuine posterior distribution, to wit the above quote: that the ABC Bayes factors favour the correct model in the simulation does not tell anything about the degree of approximation wrt the original Bayes factor. [The issue of non-consistent Bayes factors does not apply here as there is no summary statistic applied to the few observations in the data.] Further, the magnitude of the variability of the values of this Bayes factor as ε varies, from 1.3 to 9.6, mostly indicates that the numerical value is difficult to trust. (I also fail to explain the huge jump in Monte Carlo variability from 0.09 to 1.17 in Table 1.) That this form of ABC-SMC improves upon the basic ABC rejection approach is clear. However it needs to build some self-control to avoid arbitrary calibration steps and reduce the instability of the final estimates.

“The weighting function is set to be large value when the observed data and the simulated data are ‘‘close’’, small value when they are ‘‘distant’’, and constant when they are ‘‘equal’’.”

The above quote is somewhat surprising as the estimated likelihood f(xobs|xobs,θ) is naturally constant when xobs=xsim… I also failed to understand how the model intervened in the indicator function used as a default ABC kernel

## Approximate Bayesian Computation in state space models

Posted in Statistics, Travel, University life with tags , , , , , , , on October 2, 2014 by xi'an

While it took quite a while (!), with several visits by three of us to our respective antipodes, incl. my exciting trip to Melbourne and Monash University two years ago, our paper on ABC for state space models was arXived yesterday! Thanks to my coauthors, Gael Martin, Brendan McCabe, and  Worapree Maneesoonthorn,  I am very glad of this outcome and of the new perspective on ABC it produces.  For one thing, it concentrates on the selection of summary statistics from a more econometrics than usual point of view, defining asymptotic sufficiency in this context and demonstrated that both asymptotic sufficiency and Bayes consistency can be achieved when using maximum likelihood estimators of the parameters of an auxiliary model as summary statistics. In addition, the proximity to (asymptotic) sufficiency yielded by the MLE is replicated by the score vector. Using the score instead of the MLE as a summary statistics allows for huge gains in terms of speed. The method is then applied to a continuous time state space model, using as auxiliary model an augmented unscented Kalman filter. We also found in the various state space models tested therein that the ABC approach based on the marginal [likelihood] score was performing quite well, including wrt Fearnhead’s and Prangle’s (2012) approach… I like the idea of using such a generic object as the unscented Kalman filter for state space models, even when it is not a particularly accurate representation of the true model. Another appealing feature of the paper is in the connections made with indirect inference.

## this issue of Series B

Posted in Books, Statistics, Travel, University life with tags , , , , , , , , , , on September 5, 2014 by xi'an

The September issue of [JRSS] Series B I received a few days ago is of particular interest to me. (And not as an ex-co-editor since I was never involved in any of those papers!) To wit: a paper by Hani Doss and Aixin Tan on evaluating normalising constants based on MCMC output, a preliminary version I had seen at a previous JSM meeting, a paper by Nick Polson, James Scott and Jesse Windle on the Bayesian bridge, connected with Nick’s talk in Boston earlier this month, yet another paper by Ariel Kleiner, Ameet Talwalkar, Purnamrita Sarkar and Michael Jordan on the bag of little bootstraps, which presentation I heard Michael deliver a few times when he was in Paris. (Obviously, this does not imply any negative judgement on the other papers of this issue!)

For instance, Doss and Tan consider the multiple mixture estimator [my wording, the authors do not give the method a name, referring to Vardi (1985) but missing the connection with Owen and Zhou (2000)] of k ratios of normalising constants, namely

$\sum_{l=1}^k \frac{1}{n_l} \sum_{t=1}^{n_l} \dfrac{n_l g_j(x_t^l)}{\sum_{s=1}^k n_s g_s(x_t^l) z_1/z_s } \longrightarrow \dfrac{z_j}{z_1}$

where the z’s are the normalising constants and with possible different numbers of iterations of each Markov chain. An interesting starting point (that Hans Künsch had mentioned to me a while ago but that I had since then forgotten) is that the problem was reformulated by Charlie Geyer (1994) as a quasi-likelihood estimation where the ratios of all z’s relative to one reference density are the unknowns. This is doubling interesting, actually, because it restates the constant estimation problem into a statistical light and thus somewhat relates to the infamous “paradox” raised by Larry Wasserman a while ago. The novelty in the paper is (a) to derive an optimal estimator of the ratios of normalising constants in the Markov case, essentially accounting for possibly different lengths of the Markov chains, and (b) to estimate the variance matrix of the ratio estimate by regeneration arguments. A favourite tool of mine, at least theoretically as practically useful minorising conditions are hard to come by, if at all available.