Étude de la performance d’un algorithme Metropolis-Hastings avec ajustement directionnel

Les méthodes de Monte Carlo par chaîne de Markov (MCMC) sont des outils très populaires pour l’échantillonnage de lois de probabilité complexes et/ou en grandes dimensions. Étant donné leur facilité d’application, ces méthodes sont largement répandues dans plusieurs communautés scientifiques et bien certainement en statistique, particulièrement en analyse bayésienne. Depuis l’apparition de la première méthode MCMC en 1953, le nombre de ces algorithmes a considérablement augmenté et ce sujet continue d’être une aire de recherche active. Un nouvel algorithme MCMC avec ajustement directionnel a été récemment développé par Bédard et al. (IJSS, 9 :2008) et certaines de ses propriétés restent partiellement méconnues. L’objectif de ce mémoire est de tenter d’établir l’impact d’un paramètre clé de cette méthode sur la performance globale de l’approche. Un second objectif est de comparer cet algorithme à d’autres méthodes MCMC plus versatiles afin de juger de sa performance de façon relative.

Recyclage des candidats dans l'algorithme Metropolis à essais multiples

Les méthodes de Monte Carlo par chaînes de Markov (MCCM) sont des méthodes servant à échantillonner à partir de distributions de probabilité. Ces techniques se basent sur le parcours de chaînes de Markov ayant pour lois stationnaires les distributions à échantillonner. Étant donné leur facilité d’application, elles constituent une des approches les plus utilisées dans la communauté statistique, et tout particulièrement en analyse bayésienne. Ce sont des outils très populaires pour l’échantillonnage de lois de probabilité complexes et/ou en grandes dimensions. Depuis l’apparition de la première méthode MCCM en 1953 (la méthode de Metropolis, voir [10]), l’intérêt pour ces méthodes, ainsi que l’éventail d’algorithmes disponibles ne cessent de s’accroître d’une année à l’autre. Bien que l’algorithme Metropolis-Hastings (voir [8]) puisse être considéré comme l’un des algorithmes de Monte Carlo par chaînes de Markov les plus généraux, il est aussi l’un des plus simples à comprendre et à expliquer, ce qui en fait un algorithme idéal pour débuter. Il a été sujet de développement par plusieurs chercheurs. L’algorithme Metropolis à essais multiples (MTM), introduit dans la littérature statistique par [9], est considéré comme un développement intéressant dans ce domaine, mais malheureusement son implémentation est très coûteuse (en termes de temps). Récemment, un nouvel algorithme a été développé par [1]. Il s’agit de l’algorithme Metropolis à essais multiples revisité (MTM revisité), qui définit la méthode MTM standard mentionnée précédemment dans le cadre de l’algorithme Metropolis-Hastings sur un espace étendu. L’objectif de ce travail est, en premier lieu, de présenter les méthodes MCCM, et par la suite d’étudier et d’analyser les algorithmes Metropolis-Hastings ainsi que le MTM standard afin de permettre aux lecteurs une meilleure compréhension de l’implémentation de ces méthodes. Un deuxième objectif est d’étudier les perspectives ainsi que les inconvénients de l’algorithme MTM revisité afin de voir s’il répond aux attentes de la communauté statistique. Enfin, nous tentons de combattre le problème de sédentarité de l’algorithme MTM revisité, ce qui donne lieu à un tout nouvel algorithme. Ce nouvel algorithme performe bien lorsque le nombre de candidats générés à chaque itérations est petit, mais sa performance se dégrade à mesure que ce nombre de candidats croît.

Generalized Metropolis dynamics with a generalized master equation: An approach for time-independent and time-dependent Monte Carlo simulations of generalized spin systems

The extension of Boltzmann-Gibbs thermostatistics, proposed by Tsallis, introduces an additional parameter q to the inverse temperature beta. Here, we show that a previously introduced generalized Metropolis dynamics to evolve spin models is not local and does not obey the detailed energy balance. In this dynamics, locality is only retrieved for q = 1, which corresponds to the standard Metropolis algorithm. Nonlocality implies very time-consuming computer calculations, since the energy of the whole system must be reevaluated when a single spin is flipped. To circumvent this costly calculation, we propose a generalized master equation, which gives rise to a local generalized Metropolis dynamics that obeys the detailed energy balance. To compare the different critical values obtained with other generalized dynamics, we perform Monte Carlo simulations in equilibrium for the Ising model. By using short-time nonequilibrium numerical simulations, we also calculate for this model the critical temperature and the static and dynamical critical exponents as functions of q. Even for q not equal 1, we show that suitable time-evolving power laws can be found for each initial condition. Our numerical experiments corroborate the literature results when we use nonlocal dynamics, showing that short-time parameter determination works also in this case. However, the dynamics governed by the new master equation leads to different results for critical temperatures and also the critical exponents affecting universality classes. We further propose a simple algorithm to optimize modeling the time evolution with a power law, considering in a log-log plot two successive refinements.

Análise bayesiana do modelo de herança monogênica no melhoramento vegetal: um exemplo com abobrinha

A common breeding strategy is to carry out basic studies to investigate the hypothesis of a single gene controlling the trait (major gene) with or without polygenes of minor effect. In this study we used Bayesian inference to fit genetic additive-dominance models of inheritance to plant breeding experiments with multiple generations. Normal densities with different means, according to the major gene genotype, were considered in a linear model in which the design matrix of the genetic effects had unknown coefficients (which were estimated in individual basis). An actual data set from an inheritance study of partenocarpy in zucchini (Cucurbita pepo L.) was used for illustration. Model fitting included posterior probabilities for all individual genotypes. Analysis agrees with results in the literature but this approach was far more efficient than previous alternatives assuming that design matrix was known for the generations. Partenocarpy in zucchini is controlled by a major gene with important additive effect and partial dominance.

Thermodynamic origin of cooperativity in actomyosin interactions: The coupling of short-range interactions with actin bending stiffness in an Ising-like model

We present Monte Carlo simulations for a molecular motor system found in virtually all eukaryotic cells, the acto-myosin motor system, composed of a group of organic macromolecules. Cell motors were mapped to an Ising-like model, where the interaction field is transmitted through a tropomyosin polymer chain. The presence of Ca(2+) induces tropomyosin to block or unblock binding sites of the myosin motor leading to its activation or deactivation. We used the Metropolis algorithm to find the transient and the equilibrium states of the acto-myosin system composed of solvent, actin, tropomyosin, troponin, Ca(2+), and myosin-S1 at a given temperature, including the spatial configuration of tropomyosin on the actin filament surface. Our model describes the short- and long-range cooperativity during actin-myosin binding which emerges from the bending stiffness of the tropomyosin complex. We found all transition rates between the states only using the interaction energy of the constituents. The agreement between our model and experimental data also supports the recent theory of flexible tropomyosin.

Termodinâmica estatística de líquidos com o método de Monte Carlo. II. O líquido clorofórmio

Thermodynamic properties and radial distribution functions for liquid chloroform were calculated using the Monte Carlo method implemented with Metropolis algorithm in the NpT ensemble at 298 K and 1 atm. A five site model was developed to represent the chloroform molecules. A force field composed by Lennard-Jones and Coulomb potential functions was used to calculate the intermolecular energy. The partial charges needed to represent the Coulombic interactions were obtained from quantum chemical ab initio calculations. The Lennard-Jones parameters were adjusted to reproduce experimental values for density and enthalpy of vaporization for pure liquid. All thermodynamic results are in excelent agreement with experimental data. The correlation functions calculated are in good accordance with theoretical results avaliable in the literature. The free energy for solvating one chloroform molecule into its own liquid at 298 K and 1 atm was computed as an additional test of the potential model. The result obtained compares well with the experimental value. The medium effects on cis/trans convertion of a hypotetical solute in water TIP4P and chloroform solvents were also accomplished. The results obtained from this investigation are in agreement with estimates of the continuous theory of solvation.

O Modelo de Ising inomogêneo: uma interrupção contínua entre as redes quadrada e triangular.

The ferromagnetic and antiferromagnetic Ising model on a two dimensional inhomogeneous lattice characterized by two exchange constants (J1 and J2) is investigated. The lattice allows, in a continuous manner, the interpolation between the uniforme square (J2 = 0) and triangular (J2 = J1) lattices. By performing Monte Carlo simulation using the sequential Metropolis algorithm, we calculate the magnetization and the magnetic susceptibility on lattices of differents sizes. Applying the finite size scaling method through a data colappse, we obtained the critical temperatures as well as the critical exponents of the model for several values of the parameter α = J2 J1 in the [0, 1] range. The ferromagnetic case shows a linear increasing behavior of the critical temperature Tc for increasing values of α. Inwhich concerns the antiferromagnetic system, we observe a linear (decreasing) behavior of Tc, only for small values of α; in the range [0.6, 1], where frustrations effects are more pronunciated, the critical temperature Tc decays more quickly, possibly in a non-linear way, to the limiting value Tc = 0, cor-responding to the homogeneous fully frustrated antiferromagnetic triangular case.

Simulações de Monte Carlo para os modelos Ising e Blume-Capel em redes complexa

We studied the Ising model ferromagnetic as spin-1/2 and the Blume-Capel model as spin-1, > 0 on small world network, using computer simulation through the Metropolis algorithm. We calculated macroscopic quantities of the system, such as internal energy, magnetization, specific heat, magnetic susceptibility and Binder cumulant. We found for the Ising model the same result obtained by Koreans H. Hong, Beom Jun Kim and M. Y. Choi [6] and critical behavior similar Blume-Capel model

Simulação de Monte Carlo: de modelos de spin à teoria de campos na rede

Pós-graduação em Física - IFT

Study of surface free energy on a lattice-gas model applied to Liquid bridge

The pulmonary crackling and the formation of liquid bridges are problems that for centuries have been attracting the attention of scientists. In order to study these phenomena, it was developed a canonical cubic lattice-gas­ like model to explain the rupture of liquid bridges in lung airways [A. Alencar et al., 2006, PRE]. Here, we further develop this model and add entropy analysis to study thermodynamic properties, such as free energy and force. The simulations were performed using the Monte Carlo method with Metropolis algorithm. The exchange between gas and liquid particles were performed randomly according to the Kawasaki dynamics and weighted by the Boltzmann factor. Each particle, which can be solid (s), liquid (l) or gas (g), has 26 neighbors: 6 + 12 + 8, with distances 1, √2 and √3, respectively. The energy of a lattice's site m is calculated by the following expression: Em = ∑k=126 Ji(m)j(k) in witch (i, j) = g, l or s. Specifically, it was studied the surface free energy of the liquid bridge, trapped between two planes, when its height is changed. For that, was considered two methods. First, just the internal energy was calculated. Then was considered the entropy. It was fond no difference in the surface free energy between this two methods. We calculate the liquid bridge force between the two planes using the numerical surface free energy. This force is strong for small height, and decreases as the distance between the two planes, height, is increased. The liquid-gas system was also characterized studying the variation of internal energy and heat capacity with the temperature. For that, was performed simulation with the same proportion of liquid and gas particle, but different lattice size. The scale of the liquid-gas system was also studied, for low temperature, using different values to the interaction Jij.

Tethered Monte Carlo: computing the effective potential without critical slowing down

We present Tethered Monte Carlo, a simple, general purpose method of computing the effective potential of the order parameter (Helmholtz free energy). This formalism is based on a new statistical ensemble, closely related to the micromagnetic one, but with an extended configuration space (through Creutz-like demons). Canonical averages for arbitrary values of the external magnetic field are computed without additional simulations. The method is put to work in the two-dimensional Ising model, where the existence of exact results enables us to perform high precision checks. A rather peculiar feature of our implementation, which employs a local Metropolis algorithm, is the total absence, within errors, of critical slowing down for magnetic observables. Indeed, high accuracy results are presented for lattices as large as L = 1024.

Tethered Monte Carlo: computing the effective potential without critical slowing down

We present Tethered Monte Carlo, a simple, general purpose method of computing the effective potential of the order parameter (Helmholtz free energy). This formalism is based on a new statistical ensemble, closely related to the micromagnetic one, but with an extended configuration space (through Creutz-like demons). Canonical averages for arbitrary values of the external magnetic field are computed without additional simulations. The method is put to work in the two-dimensional Ising model, where the existence of exact results enables us to perform high precision checks. A rather peculiar feature of our implementation, which employs a local Metropolis algorithm, is the total absence, within errors, of critical slowing down for magnetic observables. Indeed, high accuracy results are presented for lattices as large as L = 1024.

Bayesian Inference in Large-scale Problems

Many modern applications fall into the category of "large-scale" statistical problems, in which both the number of observations n and the number of features or parameters p may be large. Many existing methods focus on point estimation, despite the continued relevance of uncertainty quantification in the sciences, where the number of parameters to estimate often exceeds the sample size, despite huge increases in the value of n typically seen in many fields. Thus, the tendency in some areas of industry to dispense with traditional statistical analysis on the basis that "n=all" is of little relevance outside of certain narrow applications. The main result of the Big Data revolution in most fields has instead been to make computation much harder without reducing the importance of uncertainty quantification. Bayesian methods excel at uncertainty quantification, but often scale poorly relative to alternatives. This conflict between the statistical advantages of Bayesian procedures and their substantial computational disadvantages is perhaps the greatest challenge facing modern Bayesian statistics, and is the primary motivation for the work presented here.

Two general strategies for scaling Bayesian inference are considered. The first is the development of methods that lend themselves to faster computation, and the second is design and characterization of computational algorithms that scale better in n or p. In the first instance, the focus is on joint inference outside of the standard problem of multivariate continuous data that has been a major focus of previous theoretical work in this area. In the second area, we pursue strategies for improving the speed of Markov chain Monte Carlo algorithms, and characterizing their performance in large-scale settings. Throughout, the focus is on rigorous theoretical evaluation combined with empirical demonstrations of performance and concordance with the theory.

One topic we consider is modeling the joint distribution of multivariate categorical data, often summarized in a contingency table. Contingency table analysis routinely relies on log-linear models, with latent structure analysis providing a common alternative. Latent structure models lead to a reduced rank tensor factorization of the probability mass function for multivariate categorical data, while log-linear models achieve dimensionality reduction through sparsity. Little is known about the relationship between these notions of dimensionality reduction in the two paradigms. In Chapter 2, we derive several results relating the support of a log-linear model to nonnegative ranks of the associated probability tensor. Motivated by these findings, we propose a new collapsed Tucker class of tensor decompositions, which bridge existing PARAFAC and Tucker decompositions, providing a more flexible framework for parsimoniously characterizing multivariate categorical data. Taking a Bayesian approach to inference, we illustrate empirical advantages of the new decompositions.

Latent class models for the joint distribution of multivariate categorical, such as the PARAFAC decomposition, data play an important role in the analysis of population structure. In this context, the number of latent classes is interpreted as the number of genetically distinct subpopulations of an organism, an important factor in the analysis of evolutionary processes and conservation status. Existing methods focus on point estimates of the number of subpopulations, and lack robust uncertainty quantification. Moreover, whether the number of latent classes in these models is even an identified parameter is an open question. In Chapter 3, we show that when the model is properly specified, the correct number of subpopulations can be recovered almost surely. We then propose an alternative method for estimating the number of latent subpopulations that provides good quantification of uncertainty, and provide a simple procedure for verifying that the proposed method is consistent for the number of subpopulations. The performance of the model in estimating the number of subpopulations and other common population structure inference problems is assessed in simulations and a real data application.

In contingency table analysis, sparse data is frequently encountered for even modest numbers of variables, resulting in non-existence of maximum likelihood estimates. A common solution is to obtain regularized estimates of the parameters of a log-linear model. Bayesian methods provide a coherent approach to regularization, but are often computationally intensive. Conjugate priors ease computational demands, but the conjugate Diaconis--Ylvisaker priors for the parameters of log-linear models do not give rise to closed form credible regions, complicating posterior inference. In Chapter 4 we derive the optimal Gaussian approximation to the posterior for log-linear models with Diaconis--Ylvisaker priors, and provide convergence rate and finite-sample bounds for the Kullback-Leibler divergence between the exact posterior and the optimal Gaussian approximation. We demonstrate empirically in simulations and a real data application that the approximation is highly accurate, even in relatively small samples. The proposed approximation provides a computationally scalable and principled approach to regularized estimation and approximate Bayesian inference for log-linear models.

Another challenging and somewhat non-standard joint modeling problem is inference on tail dependence in stochastic processes. In applications where extreme dependence is of interest, data are almost always time-indexed. Existing methods for inference and modeling in this setting often cluster extreme events or choose window sizes with the goal of preserving temporal information. In Chapter 5, we propose an alternative paradigm for inference on tail dependence in stochastic processes with arbitrary temporal dependence structure in the extremes, based on the idea that the information on strength of tail dependence and the temporal structure in this dependence are both encoded in waiting times between exceedances of high thresholds. We construct a class of time-indexed stochastic processes with tail dependence obtained by endowing the support points in de Haan's spectral representation of max-stable processes with velocities and lifetimes. We extend Smith's model to these max-stable velocity processes and obtain the distribution of waiting times between extreme events at multiple locations. Motivated by this result, a new definition of tail dependence is proposed that is a function of the distribution of waiting times between threshold exceedances, and an inferential framework is constructed for estimating the strength of extremal dependence and quantifying uncertainty in this paradigm. The method is applied to climatological, financial, and electrophysiology data.

The remainder of this thesis focuses on posterior computation by Markov chain Monte Carlo. The Markov Chain Monte Carlo method is the dominant paradigm for posterior computation in Bayesian analysis. It has long been common to control computation time by making approximations to the Markov transition kernel. Comparatively little attention has been paid to convergence and estimation error in these approximating Markov Chains. In Chapter 6, we propose a framework for assessing when to use approximations in MCMC algorithms, and how much error in the transition kernel should be tolerated to obtain optimal estimation performance with respect to a specified loss function and computational budget. The results require only ergodicity of the exact kernel and control of the kernel approximation accuracy. The theoretical framework is applied to approximations based on random subsets of data, low-rank approximations of Gaussian processes, and a novel approximating Markov chain for discrete mixture models.

Data augmentation Gibbs samplers are arguably the most popular class of algorithm for approximately sampling from the posterior distribution for the parameters of generalized linear models. The truncated Normal and Polya-Gamma data augmentation samplers are standard examples for probit and logit links, respectively. Motivated by an important problem in quantitative advertising, in Chapter 7 we consider the application of these algorithms to modeling rare events. We show that when the sample size is large but the observed number of successes is small, these data augmentation samplers mix very slowly, with a spectral gap that converges to zero at a rate at least proportional to the reciprocal of the square root of the sample size up to a log factor. In simulation studies, moderate sample sizes result in high autocorrelations and small effective sample sizes. Similar empirical results are observed for related data augmentation samplers for multinomial logit and probit models. When applied to a real quantitative advertising dataset, the data augmentation samplers mix very poorly. Conversely, Hamiltonian Monte Carlo and a type of independence chain Metropolis algorithm show good mixing on the same dataset.