## Bayesian sampling without tears

Posted in Books, Kids, R, Statistics with tags , , , , , , , , , , , , on May 24, 2022 by xi'an

Following a question on Stack Overflow trying to replicate a figure from the paper written by Alan Gelfand and Adrian Smith (1990) for The American Statistician, Bayesian sampling without tears, which precedes their historical MCMC papers, I looked at the R code produced by the OP and could not spot an issue as to why their simulation did not fit the posterior produced in the paper. Which proposes acceptance-rejection and sampling-importance-resampling as two solutions to approximately simulate from the posterior. The later being illustrated by simulations from the prior being weighted by the likelihood… The illustration is made of 3 observations from the sum of two Binomials with different success probabilities, θ¹ and θ². With a Uniform prior on both.

for (i in 1:N)
for (k in 1:3){
llh<-0
for (j in max(0,n2[k]-y[k]):min(y[k],n1[k]))
llh<-llh+choose(n1[k],j)*choose(n2[k],y[k]-j)*
theta[i,1]^j*(1-theta[i,1])^(n1[k]-j)*theta[i,2]^(y[k]-j)*
(1-theta[i,2])^(n2[k]-y[k]+j)
l[i]=l[i]*llh}


To double-check, I also wrote a Gibbs version:

theta=matrix(runif(2),nrow=T,ncol=2)
x1=rep(NA,3)
for(t in 1:(T-1)){
for(j in 1:3){
a<-max(0,n2[j]-y[j]):min(y[j],n1[j])
x1[j]=sample(a,1,
prob=choose(n1[j],a)*choose(n2[j],y[j]-a)*
theta[t,1]^a*(1-theta[t,1])^(n1[j]-a)*
theta[t,2]^(y[j]-a)*(1-theta[t,2])^(n2[j]-y[j]+a)
)}
theta[t+1,1]=rbeta(1,sum(x1)+1,sum(n1)-sum(x1)+1)
theta[t+1,2]=rbeta(1,sum(y)-sum(x1)+1,sum(n2)-sum(y)+sum(x1)+1)}


which did not show any difference with the above. Nor with the likelihood surface.

## taking advantage of the constant

Posted in Books, Kids, pictures, R, Statistics, University life with tags , , , , , , , , on May 19, 2022 by xi'an

A question from X validated had enough appeal for me to procrastinate about it for ½ an hour: what difference does it make [for simulation purposes] that a target density is properly normalised? In the continuous case, I do not see much to exploit about this knowledge, apart from the value potentially leading to a control variate (in a Gelfand and Dey 1996 spirit) and possibly to a stopping rule (by checking that the portion of the space visited so far has mass close to one, but this is more delicate than it sounds).

In a (possibly infinite) countable setting, it seems to me one gain (?) is that approximating expectations by Monte Carlo no longer requires iid simulations in the sense that once visited,  atoms need not be visited again. Self-avoiding random walks and their generalisations thus appear as a natural substitute for MC(MC) methods in this setting, provided finding unexplored atoms proves manageable. For instance, a stopping rule is always available, namely that the cumulated weight of the visited fraction of the space is close enough to one. The above picture shows a toy example on a 500 x 500 grid with 0.1% of the mass remaining at the almost invisible white dots. (In my experiment, neighbours for the random exploration were chosen at random over the grid, as I assumed no global information was available about the repartition over the grid either of mass function or of the function whose expectation was seeked.)

## rethinking the ESS published!

Posted in Statistics with tags , , , , , , , , on May 3, 2022 by xi'an

Our paper Rethinking the Effective Sample Size, with Victor Elvira (the driving force behind the paper!) and Luca Martino, has now been published in the International Statistical Review! As discussed earlier on this blog, we wanted to re-evaluate the pros and cons of the effective sample size (ESS), as a tool assessing the quality [or lack thereof] of a Monte Carlo approximation. It is particularly exploited in the specific context of importance sampling. Following a 1992 construction by Augustine Kong, his approximation has been widely used in the last 25 years, in part due to its simplicity as a practical rule of thumb. However, we show in this paper that the assumptions made in the derivation of this approximation make it difficult to consider it as a reasonable approximation of the ESS. Note that this reevaluation does not cover the use of ESS for Markov chain Monte Carlo algorithms, although there would also be much to tell about it..!

## efficiency of normalising over discrete parameters

Posted in Statistics with tags , , , , , , , , , on May 1, 2022 by xi'an

Yesterday, I noticed a new arXival entitled Investigating the efficiency of marginalising over discrete parameters in Bayesian computations written by Wen Wang and coauthors. The paper is actually comparing the simulation of a Gibbs sampler with an Hamiltonian Monte Carlo approach on Gaussian mixtures, when including and excluding latent variables, respectively. The authors missed the opposite marginalisation when the parameters are integrated.

While marginalisation requires substantial mathematical effort, folk wisdom in the Stan community suggests that fitting models with marginalisation is more efficient than using Gibbs sampling.

The comparison is purely experimental, though, which means it depends on the simulated data, the sample size, the prior selection, and of course the chosen algorithms. It also involves the [mostly] automated [off-the-shelf] choices made in the adopted software, JAGS and Stan. The outcome is only evaluated through ESS and the (old) R statistic. Which all depend on the parameterisation. But evacuates the label switching problem by imposing an ordering on the Gaussian means, which may have a different impact on marginalised and unmarginalised models. All in all, there is not much one can conclude about this experiment since the parameter values beyond the simulated data seem to impact the performances much more than the type of algorithm one implements.

## null recurrent = zero utility?

Posted in Books, R, Statistics with tags , , , , , , , on April 28, 2022 by xi'an

The stability result that the ratio

$\dfrac{\sum^T_{t=1} f(\theta^{(t)})}{\sum^T_{t=1} g(\theta^{(t)})}\qquad(1)$

converges holds for a Harris π-null-recurrent Markov chain for all functions f,g in L¹(π) [Meyn & Tweedie, 1993, Theorem 17.3.2] is rather fascinating. However, it is unclear it can be useful in simulation environments, as for the integral priors we have been studying over the years with Juan Antonio Cano and Diego Salmeron Martinez. Above, the result of an experiment where I simulated a Markov chain as a Normal random walk in dimension one, hence a Harris π-null-recurrent Markov chain for the Lebesgue measure λ, and monitored the stabilisation of the ratio (1) when using two densities for f and g,  to its expected value (1, shown by a red horizontal line). There is quite a variability in the outcome (repeated 100 times),  but the most intriguing is the quick stabilisation of most cumulated averages to values different from 1. Even longer runs display this feature

which I would blame on the excursions of the random walk far away from the central regions for both f and g, that is on long sequences where zeroes keep being added to numerator and denominators in (1). As far as integral approximation is concerned, this is not very helpful!