Realized Matrix-Exponential Stochastic Volatility with Asymmetry, Long Memory and Spillovers

The paper develops a novel realized matrix-exponential stochastic volatility model of multivariate returns and realized covariances that incorporates asymmetry and long memory (hereafter the RMESV-ALM model). The matrix exponential transformation guarantees the positivedefiniteness of the dynamic covariance matrix. The contribution of the paper ties in with Robert Basmann’s seminal work in terms of the estimation of highly non-linear model specifications (“Causality tests and observationally equivalent representations of econometric models”, Journal of Econometrics, 1988, 39(1-2), 69–104), especially for developing tests for leverage and spillover effects in the covariance dynamics. Efficient importance sampling is used to maximize the likelihood function of RMESV-ALM, and the finite sample properties of the quasi-maximum likelihood estimator of the parameters are analysed. Using high frequency data for three US financial assets, the new model is estimated and evaluated. The forecasting performance of the new model is compared with a novel dynamic realized matrix-exponential conditional covariance model. The volatility and co-volatility spillovers are examined via the news impact curves and the impulse response functions from returns to volatility and co-volatility.

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.

Essays on Microeconometrics

My dissertation has three chapters which develop and apply microeconometric tech- niques to empirically relevant problems. All the chapters examines the robustness issues (e.g., measurement error and model misspecification) in the econometric anal- ysis. The first chapter studies the identifying power of an instrumental variable in the nonparametric heterogeneous treatment effect framework when a binary treat- ment variable is mismeasured and endogenous. I characterize the sharp identified set for the local average treatment effect under the following two assumptions: (1) the exclusion restriction of an instrument and (2) deterministic monotonicity of the true treatment variable in the instrument. The identification strategy allows for general measurement error. Notably, (i) the measurement error is nonclassical, (ii) it can be endogenous, and (iii) no assumptions are imposed on the marginal distribution of the measurement error, so that I do not need to assume the accuracy of the measure- ment. Based on the partial identification result, I provide a consistent confidence interval for the local average treatment effect with uniformly valid size control. I also show that the identification strategy can incorporate repeated measurements to narrow the identified set, even if the repeated measurements themselves are endoge- nous. Using the the National Longitudinal Study of the High School Class of 1972, I demonstrate that my new methodology can produce nontrivial bounds for the return to college attendance when attendance is mismeasured and endogenous.

The second chapter, which is a part of a coauthored project with Federico Bugni, considers the problem of inference in dynamic discrete choice problems when the structural model is locally misspecified. We consider two popular classes of estimators for dynamic discrete choice models: K-step maximum likelihood estimators (K-ML) and K-step minimum distance estimators (K-MD), where K denotes the number of policy iterations employed in the estimation problem. These estimator classes include popular estimators such as Rust (1987)’s nested fixed point estimator, Hotz and Miller (1993)’s conditional choice probability estimator, Aguirregabiria and Mira (2002)’s nested algorithm estimator, and Pesendorfer and Schmidt-Dengler (2008)’s least squares estimator. We derive and compare the asymptotic distributions of K- ML and K-MD estimators when the model is arbitrarily locally misspecified and we obtain three main results. In the absence of misspecification, Aguirregabiria and Mira (2002) show that all K-ML estimators are asymptotically equivalent regardless of the choice of K. Our first result shows that this finding extends to a locally misspecified model, regardless of the degree of local misspecification. As a second result, we show that an analogous result holds for all K-MD estimators, i.e., all K- MD estimator are asymptotically equivalent regardless of the choice of K. Our third and final result is to compare K-MD and K-ML estimators in terms of asymptotic mean squared error. Under local misspecification, the optimally weighted K-MD estimator depends on the unknown asymptotic bias and is no longer feasible. In turn, feasible K-MD estimators could have an asymptotic mean squared error that is higher or lower than that of the K-ML estimators. To demonstrate the relevance of our asymptotic analysis, we illustrate our findings using in a simulation exercise based on a misspecified version of Rust (1987) bus engine problem.

The last chapter investigates the causal effect of the Omnibus Budget Reconcil- iation Act of 1993, which caused the biggest change to the EITC in its history, on unemployment and labor force participation among single mothers. Unemployment and labor force participation are difficult to define for a few reasons, for example, be- cause of marginally attached workers. Instead of searching for the unique definition for each of these two concepts, this chapter bounds unemployment and labor force participation by observable variables and, as a result, considers various competing definitions of these two concepts simultaneously. This bounding strategy leads to partial identification of the treatment effect. The inference results depend on the construction of the bounds, but they imply positive effect on labor force participa- tion and negligible effect on unemployment. The results imply that the difference- in-difference result based on the BLS definition of unemployment can be misleading

due to misclassification of unemployment.

Multivariate Appoximations to Portfolio Return Distribution

This article proposes a three-step procedure to estimate portfolio return distributions under the multivariate Gram-Charlier (MGC) distribution. The method combines quasi maximum likelihood (QML) estimation for conditional means and variances and the method of moments (MM) estimation for the rest of the density parameters, including the correlation coefficients. The procedure involves consistent estimates even under density misspecification and solves the so-called ‘curse of dimensionality’ of multivariate modelling. Furthermore, the use of a MGC distribution represents a flexible and general approximation to the true distribution of portfolio returns and accounts for all its empirical regularities. An application of such procedure is performed for a portfolio composed of three European indices as an illustration. The MM estimation of the MGC (MGC-MM) is compared with the traditional maximum likelihood of both the MGC and multivariate Student’s t (benchmark) densities. A simulation on Value-at-Risk (VaR) performance for an equally weighted portfolio at 1% and 5% confidence indicates that the MGC-MM method provides reasonable approximations to the true empirical VaR. Therefore, the procedure seems to be a useful tool for risk managers and practitioners.

A timescale for diatom evolution based on four molecular markers: reassessment of ghost lineages and major steps defining diatom evolution.

ABSTRACT. – Phylogenies and molecular clocks of the diatoms have largely been inferred from SSU rDNA sequences. A new phylogeny of diatoms was estimated using four gene markers SSU and LSU rDNA rbcL and psbA (total 4352 bp) with 42 diatom species. The four gene trees analysed with a maximum likelihood (ML) and Baysian (BI) analysis recovered a monophyletic origin of the new diatom classes with high bootstrap support, which has been controversial with single gene markers using single outgroups and alignments that do not take secondary structure of the SSU gene into account. The divergence time of the classes were calculated from a ML tree in the MultliDiv Time program using a Bayesian estimation allowing for simultaneous constraints from the fossil record and varying rates of molecular evolution of different branches in the phylogenetic tree. These divergence times are generally in agreement with those proposed by other clocks using single genes with the exception that the pennates appear much earlier and suggest a longer Cretaceous fossil record that has yet to be sampled. Ghost lineages (i.e. the discrepancy between first appearance (FA) and molecular clock age of origin from an extant taxon) were revealed in the pennate lineage, whereas those ghost lineages in the centric lineages previously reported by others are reviewed and referred to earlier literature.

A timescale for diatom evolution based on four molecular markers: reassessment of ghost lineages and major steps defining diatom evolution.

ABSTRACT. – Phylogenies and molecular clocks of the diatoms have largely been inferred from SSU rDNA sequences. A new phylogeny of diatoms was estimated using four gene markers SSU and LSU rDNA rbcL and psbA (total 4352 bp) with 42 diatom species. The four gene trees analysed with a maximum likelihood (ML) and Baysian (BI) analysis recovered a monophyletic origin of the new diatom classes with high bootstrap support, which has been controversial with single gene markers using single outgroups and alignments that do not take secondary structure of the SSU gene into account. The divergence time of the classes were calculated from a ML tree in the MultliDiv Time program using a Bayesian estimation allowing for simultaneous constraints from the fossil record and varying rates of molecular evolution of different branches in the phylogenetic tree. These divergence times are generally in agreement with those proposed by other clocks using single genes with the exception that the pennates appear much earlier and suggest a longer Cretaceous fossil record that has yet to be sampled. Ghost lineages (i.e. the discrepancy between first appearance (FA) and molecular clock age of origin from an extant taxon) were revealed in the pennate lineage, whereas those ghost lineages in the centric lineages previously reported by others are reviewed and referred to earlier literature.

Directional genetic differentiation and relative migration

Understanding the population structure and patterns of gene flow within species is of fundamental importance to the study of evolution. In the fields of population and evolutionary genetics, measures of genetic differentiation are commonly used to gather this information. One potential caveat is that these measures assume gene flow to be symmetric. However, asymmetric gene flow is common in nature, especially in systems driven by physical processes such as wind or water currents. As information about levels of asymmetric gene flow among populations is essential for the correct interpretation of the distribution of contemporary genetic diversity within species, this should not be overlooked. To obtain information on asymmetric migration patterns from genetic data, complex models based on maximum-likelihood or Bayesian approaches generally need to be employed, often at great computational cost. Here, a new simpler and more efficient approach for understanding gene flow patterns is presented. This approach allows the estimation of directional components of genetic divergence between pairs of populations at low computational effort, using any of the classical or modern measures of genetic differentiation. These directional measures of genetic differentiation can further be used to calculate directional relative migration and to detect asymmetries in gene flow patterns. This can be done in a user-friendly web application called divMigrate-online introduced in this study. Using simulated data sets with known gene flow regimes, we demonstrate that the method is capable of resolving complex migration patterns under a range of study designs.

Modelos autorregressivos com valores inteiros não negativos : aplicação em séries económicas de Cabo Verde

No estudo de séries temporais, os processos estocásticos usuais assumem que as distribuições marginais são contínuas e, em geral, não são adequados para modelar séries de contagem, pois as suas características não lineares colocam alguns problemas estatísticos, principalmente na estimação dos parâmetros. Assim, investigou-se metodologias apropriadas de análise e modelação de séries com distribuições marginais discretas. Neste contexto, Al-Osh and Alzaid (1987) e McKenzie (1988) introduziram na literatura a classe dos modelos autorregressivos com valores inteiros não negativos, os processos INAR. Estes modelos têm sido frequentemente tratados em artigos científicos ao longo das últimas décadas, pois a sua importância nas aplicações em diversas áreas do conhecimento tem despertado um grande interesse no seu estudo. Neste trabalho, após uma breve revisão sobre séries temporais e os métodos clássicos para a sua análise, apresentamos os modelos autorregressivos de valores inteiros não negativos de primeira ordem INAR (1) e a sua extensão para uma ordem p, as suas propriedades e alguns métodos de estimação dos parâmetros nomeadamente, o método de Yule-Walker, o método de Mínimos Quadrados Condicionais (MQC), o método de Máxima Verosimilhança Condicional (MVC) e o método de Quase Máxima Verosimilhança (QMV). Apresentamos também um critério automático de seleção de ordem para modelos INAR, baseado no Critério de Informação de Akaike Corrigido, AICC, um dos critérios usados para determinar a ordem em modelos autorregressivos, AR. Finalmente, apresenta-se uma aplicação da metodologia dos modelos INAR em dados reais de contagem relativos aos setores dos transportes marítimos e atividades de seguros de Cabo Verde.

Risque de prix et décisions de production et d'exportation : le cas de l'agriculture au Québec

Cette thèse porte sur l’effet du risque de prix sur la décision des agriculteurs et les transformateurs québécois. Elle se divise en trois chapitres. Le premier chapitre revient sur la littérature. Le deuxième chapitre examine l’effet du risque de prix sur la production de trois produits, à savoir le maïs grain, la viande de porc et la viande d’agneau dans la province Québec. Le dernier chapitre est centré sur l’analyse de changement des préférences du transformateur québécois de porc pour ce qui est du choix de marché. Le premier chapitre vise à montrer l’importance de l’effet du risque du prix sur la quantité produite par les agriculteurs, tel que mis en évidence par la littérature. En effet, la littérature révèle l’importance du risque de prix à l’exportation sur le commerce international. Le deuxième chapitre est consacré à l’étude des facteurs du risque (les anticipations des prix et la volatilité des prix) dans la fonction de l’offre. Un modèle d’hétéroscédasticité conditionnelle autorégressive généralisée (GARCH) est utilisé afin de modéliser ces facteurs du risque. Les paramètres du modèle sont estimés par la méthode de l’Information Complète Maximum Vraisemblance (FIML). Les résultats empiriques montrent l’effet négatif de la volatilité du prix sur la production alors que la prévisibilité des prix a un effet positif sur la quantité produite. Comme attendu, nous constatons que l’application du programme d’assurance-stabilisation des revenus agricoles (ASRA) au Québec induit une plus importante sensibilité de l’offre par rapport au prix effectif (le prix incluant la compensation de l’ASRA) que par rapport au prix du marché. Par ailleurs, l’offre est moins sensible au prix des intrants qu’au prix de l’output. La diminution de l’aversion au risque de producteur est une autre conséquence de l’application de ce programme. En outre, l’estimation de la prime marginale relative au risque révèle que le producteur du maïs est le producteur le moins averse au risque (comparativement à celui de porc ou d’agneau). Le troisième chapitre consiste en l’analyse du changement de préférence du transformateur québécois du porc pour ce qui est du choix de marché. Nous supposons que le transformateur a la possibilité de fournir les produits sur deux marchés : étranger et local. Le modèle théorique explique l’offre relative comme étant une fonction à la fois d’anticipation relative et de volatilité relative des prix. Ainsi, ce modèle révèle que la sensibilité de l’offre relative par rapport à la volatilité relative de prix dépend de deux facteurs : d’une part, la part de l’exportation dans la production totale et d’autre part, l’élasticité de substitution entre les deux marchés. Un modèle à correction d’erreurs est utilisé lors d’estimation des paramètres du modèle. Les résultats montrent l’effet positif et significatif de l’anticipation relative du prix sur l’offre relative à court terme. Ces résultats montrent donc qu’une hausse de la volatilité du prix sur le marché étranger par rapport à celle sur le marché local entraine une baisse de l’offre relative sur le marché étranger à long terme. De plus, selon les résultats, les marchés étranger et local sont plus substituables à long terme qu’à court terme.

Methods for Hypothesis Testing in Animal Carcinogenicity Experiments

Thesis (Ph.D.)--University of Washington, 2016-08

Statistical modeling of protein lysate array data

The protein lysate array is an emerging technology for quantifying the protein concentration ratios in multiple biological samples. It is gaining popularity, and has the potential to answer questions about post-translational modifications and protein pathway relationships. Statistical inference for a parametric quantification procedure has been inadequately addressed in the literature, mainly due to two challenges: the increasing dimension of the parameter space and the need to account for dependence in the data. Each chapter of this thesis addresses one of these issues. In Chapter 1, an introduction to the protein lysate array quantification is presented, followed by the motivations and goals for this thesis work. In Chapter 2, we develop a multi-step procedure for the Sigmoidal models, ensuring consistent estimation of the concentration level with full asymptotic efficiency. The results obtained in this chapter justify inferential procedures based on large-sample approximations. Simulation studies and real data analysis are used to illustrate the performance of the proposed method in finite-samples. The multi-step procedure is simpler in both theory and computation than the single-step least squares method that has been used in current practice. In Chapter 3, we introduce a new model to account for the dependence structure of the errors by a nonlinear mixed effects model. We consider a method to approximate the maximum likelihood estimator of all the parameters. Using the simulation studies on various error structures, we show that for data with non-i.i.d. errors the proposed method leads to more accurate estimates and better confidence intervals than the existing single-step least squares method.

Modelos autorregressivos com valores inteiros não negativos : aplicação em séries económicas de Cabo Verde

No estudo de séries temporais, os processos estocásticos usuais assumem que as distribuições marginais são contínuas e, em geral, não são adequados para modelar séries de contagem, pois as suas características não lineares colocam alguns problemas estatísticos, principalmente na estimação dos parâmetros. Assim, investigou-se metodologias apropriadas de análise e modelação de séries com distribuições marginais discretas. Neste contexto, Al-Osh and Alzaid (1987) e McKenzie (1988) introduziram na literatura a classe dos modelos autorregressivos com valores inteiros não negativos, os processos INAR. Estes modelos têm sido frequentemente tratados em artigos científicos ao longo das últimas décadas, pois a sua importância nas aplicações em diversas áreas do conhecimento tem despertado um grande interesse no seu estudo. Neste trabalho, após uma breve revisão sobre séries temporais e os métodos clássicos para a sua análise, apresentamos os modelos autorregressivos de valores inteiros não negativos de primeira ordem INAR (1) e a sua extensão para uma ordem p, as suas propriedades e alguns métodos de estimação dos parâmetros nomeadamente, o método de Yule-Walker, o método de Mínimos Quadrados Condicionais (MQC), o método de Máxima Verosimilhança Condicional (MVC) e o método de Quase Máxima Verosimilhança (QMV). Apresentamos também um critério automático de seleção de ordem para modelos INAR, baseado no Critério de Informação de Akaike Corrigido, AICC, um dos critérios usados para determinar a ordem em modelos autorregressivos, AR. Finalmente, apresenta-se uma aplicação da metodologia dos modelos INAR em dados reais de contagem relativos aos setores dos transportes marítimos e atividades de seguros de Cabo Verde.

Essays in Financial Econometrics

The first paper sheds light on the informational content of high frequency data and daily data. I assess the economic value of the two family models comparing their performance in forecasting asset volatility through the Value at Risk metric. In running the comparison this paper introduces two key assumptions: jumps in prices and leverage effect in volatility dynamics. Findings suggest that high frequency data models do not exhibit a superior performance over daily data models. In the second paper, building on Majewski et al. (2015), I propose an affine-discrete time model, labeled VARG-J, which is characterized by a multifactor volatility specification. In the VARG-J model volatility experiences periods of extreme movements through a jump factor modeled as an Autoregressive Gamma Zero process. The estimation under historical measure is done by quasi-maximum likelihood and the Extended Kalman Filter. This strategy allows to filter out both volatility factors introducing a measurement equation that relates the Realized Volatility to latent volatility. The risk premia parameters are calibrated using call options written on S&P500 Index. The results clearly illustrate the important contribution of the jump factor in the pricing performance of options and the economic significance of the volatility jump risk premia. In the third paper, I analyze whether there is empirical evidence of contagion at the bank level, measuring the direction and the size of contagion transmission between European markets. In order to understand and quantify the contagion transmission on banking market, I estimate the econometric model by Aït-Sahalia et al. (2015) in which contagion is defined as the within and between countries transmission of shocks and asset returns are directly modeled as a Hawkes jump diffusion process. The empirical analysis indicates that there is a clear evidence of contagion from Greece to European countries as well as self-contagion in all countries.

Censored Linear Regression Models For Irregularly Observed Longitudinal Data Using The Multivariate-t Distribution.

In acquired immunodeficiency syndrome (AIDS) studies it is quite common to observe viral load measurements collected irregularly over time. Moreover, these measurements can be subjected to some upper and/or lower detection limits depending on the quantification assays. A complication arises when these continuous repeated measures have a heavy-tailed behavior. For such data structures, we propose a robust structure for a censored linear model based on the multivariate Student's t-distribution. To compensate for the autocorrelation existing among irregularly observed measures, a damped exponential correlation structure is employed. An efficient expectation maximization type algorithm is developed for computing the maximum likelihood estimates, obtaining as a by-product the standard errors of the fixed effects and the log-likelihood function. The proposed algorithm uses closed-form expressions at the E-step that rely on formulas for the mean and variance of a truncated multivariate Student's t-distribution. The methodology is illustrated through an application to an Human Immunodeficiency Virus-AIDS (HIV-AIDS) study and several simulation studies.