Archive for random forests

selecting summary statistics [a tale of two distances]

Posted in Books, Statistics with tags , , , , , , , , , , , , , , on May 23, 2019 by xi'an

As Jonathan Harrison came to give a seminar in Warwick [which I could not attend], it made me aware of his paper with Ruth Baker on the selection of summaries in ABC. The setting is an ABC-SMC algorithm and it relates with Fearnhead and Prangle (2012), Barnes et al. (2012), our own random forest approach, the neural network version of Papamakarios and Murray (2016), and others. The notion here is to seek the optimal weights of different summary statistics in the tolerance distance, towards a maximization of a distance (Hellinger) between prior and ABC posterior (Wasserstein also comes to mind!). A sort of dual of the least informative prior. Estimated by a k-nearest neighbour version [based on samples from the prior and from the ABC posterior] I had never seen before. I first did not get how this k-nearest neighbour distance could be optimised in the weights since the posterior sample was already generated and (SMC) weighted, but the ABC sample can be modified by changing the [tolerance] distance weights and the resulting Hellinger distance optimised this way. (There are two distances involved, in case the above description is too murky!)

“We successfully obtain an informative unbiased posterior.”

The paper spends a significant while in demonstrating that the k-nearest neighbour estimator converges and much less on the optimisation procedure itself, which seems like a real challenge to me when facing a large number of particles and a high enough dimension (in the number of statistics). (In the examples, the size of the summary is 1 (where does the weight matter?), 32, 96, 64, with 5 10⁴, 5 10⁴, 5 10³ and…10 particles, respectively.) The authors address the issue, though, albeit briefly, by mentioning that, for the same overall computation time, the adaptive weight ABC is indeed further from the prior than a regular ABC with uniform weights [rather than weighted by the precisions]. They also argue that down-weighting some components is akin to selecting a subset of summaries, but I beg to disagree with this statement as the weights are never exactly zero, as far as I can see, hence failing to fight the curse of dimensionality. Some LASSO version could implement this feature.

over-confident about mis-specified models?

Posted in Books, pictures, Statistics, University life with tags , , , , , , , , , , , , , , , on April 30, 2019 by xi'an

Ziheng Yang and Tianqui Zhu published a paper in PNAS last year that criticises Bayesian posterior probabilities used in the comparison of models under misspecification as “overconfident”. The paper is written from a phylogeneticist point of view, rather than from a statistician’s perspective, as shown by the Editor in charge of the paper [although I thought that, after Steve Fienberg‘s intervention!, a statistician had to be involved in a submission relying on statistics!] a paper , but the analysis is rather problematic, at least seen through my own lenses… With no statistical novelty, apart from looking at the distribution of posterior probabilities in toy examples. The starting argument is that Bayesian model comparison is often reporting posterior probabilities in favour of a particular model that are close or even equal to 1.

“The Bayesian method is widely used to estimate species phylogenies using molecular sequence data. While it has long been noted to produce spuriously high posterior probabilities for trees or clades, the precise reasons for this over confidence are unknown. Here we characterize the behavior of Bayesian model selection when the compared models are misspecified and demonstrate that when the models are nearly equally wrong, the method exhibits unpleasant polarized behaviors,supporting one model with high confidence while rejecting others. This provides an explanation for the empirical observation of spuriously high posterior probabilities in molecular phylogenetics.”

The paper focus on the behaviour of posterior probabilities to strongly support a model against others when the sample size is large enough, “even when” all models are wrong, the argument being apparently that the correct output should be one of equal probability between models, or maybe a uniform distribution of these model probabilities over the probability simplex. Why should it be so?! The construction of the posterior probabilities is based on a meta-model that assumes the generating model to be part of a list of mutually exclusive models. It does not account for cases where “all models are wrong” or cases where “all models are right”. The reported probability is furthermore epistemic, in that it is relative to the measure defined by the prior modelling, not to a promise of a frequentist stabilisation in a ill-defined asymptotia. By which I mean that a 99.3% probability of model M¹ being “true”does not have a universal and objective meaning. (Moderation note: the high polarisation of posterior probabilities was instrumental in our investigation of model choice with ABC tools and in proposing instead error rates in ABC random forests.)

The notion that two models are equally wrong because they are both exactly at the same Kullback-Leibler distance from the generating process (when optimised over the parameter) is such a formal [or cartoonesque] notion that it does not make much sense. There is always one model that is slightly closer and eventually takes over. It is also bizarre that the argument does not account for the complexity of each model and the resulting (Occam’s razor) penalty. Even two models with a single parameter are not necessarily of intrinsic dimension one, as shown by DIC. And thus it is not a surprise if the posterior probability mostly favours one versus the other. In any case, an healthily sceptic approach to Bayesian model choice means looking at the behaviour of the procedure (Bayes factor, posterior probability, posterior predictive, mixture weight, &tc.) under various assumptions (model M¹, M², &tc.) to calibrate the numerical value, rather than taking it at face value. By which I do not mean a frequentist evaluation of this procedure. Actually, it is rather surprising that the authors of the PNAS paper do not jump on the case when the posterior probability of model M¹ say is uniformly distributed, since this would be a perfect setting when the posterior probability is a p-value. (This is also what happens to the bootstrapped version, see the last paragraph of the paper on p.1859, the year Darwin published his Origin of Species.)

ABC variable selection

Posted in Books, Mountains, pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , on July 18, 2018 by xi'an

Prior to the ISBA 2018 meeting, Yi Liu, Veronika Ročková, and Yuexi Wang arXived a paper on relying ABC for finding relevant variables, which is a very original approach in that ABC is not as much the object as it is a tool. And which Veronika considered during her Susie Bayarri lecture at ISBA 2018. In other words, it is not about selecting summary variables for running ABC but quite the opposite, selecting variables in a non-linear model through an ABC step. I was going to separate the two selections into algorithmic and statistical selections, but it is more like projections in the observation and covariate spaces. With ABC still providing an appealing approach to approximate the marginal likelihood. Now, one may wonder at the relevance of ABC for variable selection, aka model choice, given our warning call of a few years ago. But the current paper does not require low-dimension summary statistics, hence avoids the difficulty with the “other” Bayes factor.

In the paper, the authors consider a spike-and… forest prior!, where the Bayesian CART selection of active covariates proceeds through a regression tree, selected covariates appearing in the tree and others not appearing. With a sparsity prior on the tree partitions and this new ABC approach to select the subset of active covariates. A specific feature is in splitting the data, one part to learn about the regression function, simulating from this function and comparing with the remainder of the data. The paper further establishes that ABC Bayesian Forests are consistent for variable selection.

“…we observe a curious empirical connection between π(θ|x,ε), obtained with ABC Bayesian Forests  and rescaled variable importances obtained with Random Forests.”

The difference with our ABC-RF model choice paper is that we select summary statistics [for classification] rather than covariates. For instance, in the current paper, simulation of pseudo-data will depend on the selected subset of covariates, meaning simulating a model index, and then generating the pseudo-data, acceptance being a function of the L² distance between data and pseudo-data. And then relying on all ABC simulations to find which variables are in more often than not to derive the median probability model of Barbieri and Berger (2004). Which does not work very well if implemented naïvely. Because of the immense size of the model space, it is quite hard to find pseudo-data close to actual data, resulting in either very high tolerance or very low acceptance. The authors get over this difficulty by a neat device that reminds me of fractional or intrinsic (pseudo-)Bayes factors in that the dataset is split into two parts, one that learns about the posterior given the model index and another one that simulates from this posterior to compare with the left-over data. Bringing simulations closer to the data. I do not remember seeing this trick before in ABC settings, but it is very neat, assuming the small data posterior can be simulated (which may be a fundamental reason for the trick to remain unused!). Note that the split varies at each iteration, which means there is no impact of ordering the observations.

machine learning methods are useful for ABC [or my first PCI Evol Biol!]

Posted in Books, Kids, pictures, Statistics, University life with tags , , , , , , on November 23, 2017 by xi'an

While I am still working on setting a PCI [peer community in] Comput Stats, having secure sponsorship of some societies (ASA, KSS, RSS, SFdS, and hopefully ISBA), my coauthors Jean-Michel Marin and Louis Raynal submitted our paper ABC random forests for Bayesian parameter inference to PCI Evol Biol. And after a few months of review, including a revision accounting for the reviewers’ requests, our paper stood the test and the recommendation by Michael Blum and Dennis Prangle got published there. Great news, and hopefully helpful for our submission within the coming days!

Astrostatistics school

Posted in Mountains, pictures, R, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , , , , on October 17, 2017 by xi'an

What a wonderful week at the Astrostat [Indian] summer school in Autrans! The setting was superb, on the high Vercors plateau overlooking both Grenoble [north] and Valence [west], with the colours of the Fall at their brightest on the foliage of the forests rising on both sides of the valley and a perfect green on the fields at the centre, with sun all along, sharp mornings and warm afternoons worthy of a late Indian summer, too many running trails [turning into X country ski trails in the Winter] to contemplate for a single week [even with three hours of running over two days], many climbing sites on the numerous chalk cliffs all around [but a single afternoon for that, more later in another post!]. And of course a group of participants eager to learn about Bayesian methodology and computational algorithms, from diverse [astronomy, cosmology and more] backgrounds, trainings and countries. I was surprised at the dedication of the participants travelling all the way from Chile, Péru, and Hong Kong for the sole purpose of attending the school. David van Dyk gave the first part of the school on Bayesian concepts and MCMC methods, Roberto Trotta the second part on Bayesian model choice and hierarchical models, and myself a third part on, surprise, surprise!, approximate Bayesian computation. Plus practicals on R.

As it happens Roberto had to cancel his participation and I turned for a session into Christian Roberto, presenting his slides in the most objective possible fashion!, as a significant part covered nested sampling and Savage-Dickey ratios, not exactly my favourites for estimating constants. David joked that he was considering postponing his flight to see me talk about these, but I hope I refrained from engaging into controversy and criticisms… If anything because this was not of interest for the participants. Indeed when I started presenting ABC through what I thought was a pedestrian example, namely Rasmus Baath’s socks, I found that the main concern was not running an MCMC sampler or a substitute ABC algorithm but rather an healthy questioning of the construction of the informative prior in that artificial setting, which made me quite glad I had planned to cover this example rather than an advanced model [as, e.g., one of those covered in the packages abc, abctools, or abcrf]. Because it generated those questions about the prior [why a Negative Binomial? why these hyperparameters? &tc.] and showed how programming ABC turned into a difficult exercise even in this toy setting. And while I wanted to give my usual warning about ABC model choice and argue for random forests as a summary selection tool, I feel I should have focussed instead on another example, as this exercise brings out so clearly the conceptual difficulties with what is taught. Making me quite sorry I had to leave one day earlier. [As did missing an extra run!] Coming back by train through the sunny and grape-covered slopes of Burgundy hills was an extra reward [and no one in the train commented about the local cheese travelling in my bag!]