Results 1  10
of
14
A tutorial on MM algorithms
 Amer. Statist
, 2004
"... Most problems in frequentist statistics involve optimization of a function such as a likelihood or a sum of squares. EM algorithms are among the most effective algorithms for maximum likelihood estimation because they consistently drive the likelihood uphill by maximizing a simple surrogate function ..."
Abstract

Cited by 67 (3 self)
 Add to MetaCart
Most problems in frequentist statistics involve optimization of a function such as a likelihood or a sum of squares. EM algorithms are among the most effective algorithms for maximum likelihood estimation because they consistently drive the likelihood uphill by maximizing a simple surrogate function for the loglikelihood. Iterative optimization of a surrogate function as exemplified by an EM algorithm does not necessarily require missing data. Indeed, every EM algorithm is a special case of the more general class of MM optimization algorithms, which typically exploit convexity rather than missing data in majorizing or minorizing an objective function. In our opinion, MM algorithms deserve to part of the standard toolkit of professional statisticians. The current article explains the principle behind MM algorithms, suggests some methods for constructing them, and discusses some of their attractive features. We include numerous examples throughout the article to illustrate the concepts described. In addition to surveying previous work on MM algorithms, this article introduces some new material on constrained optimization and standard error estimation. Key words and phrases: constrained optimization, EM algorithm, majorization, minorization, NewtonRaphson 1 1
Statistical inference for discretely observed Markov jump processes
 Journal of the Royal Statistical Society: Series B (Statistical Methodology
"... Likelihood inference for discretely observed Markov jump processes with finite state space is investigated. The existence and uniqueness of the maximum likelihood estimator of the intensity matrix are investigated. This topic is closely related to the imbedding problem for Markov chains. It is demon ..."
Abstract

Cited by 19 (3 self)
 Add to MetaCart
Likelihood inference for discretely observed Markov jump processes with finite state space is investigated. The existence and uniqueness of the maximum likelihood estimator of the intensity matrix are investigated. This topic is closely related to the imbedding problem for Markov chains. It is demonstrated that the maximum likelihood estimator can be found either by the EMalgorithm or by a Markov chain Monte Carlo procedure. When the maximum likelihood estimator does not exist, an estimator can be obtained by using a penalized likelihood function or by the MCMCprocedure with a suitable prior. The theory is illustrated by a simulation study.
A survey of Monte Carlo algorithms for maximizing the likelihood of a twostage hierarchical model
, 2001
"... Likelihood inference with hierarchical models is often complicated by the fact that the likelihood function involves intractable integrals. Numerical integration (e.g. quadrature) is an option if the dimension of the integral is low but quickly becomes unreliable as the dimension grows. An alternati ..."
Abstract

Cited by 10 (4 self)
 Add to MetaCart
Likelihood inference with hierarchical models is often complicated by the fact that the likelihood function involves intractable integrals. Numerical integration (e.g. quadrature) is an option if the dimension of the integral is low but quickly becomes unreliable as the dimension grows. An alternative approach is to approximate the intractable integrals using Monte Carlo averages. Several dierent algorithms based on this idea have been proposed. In this paper we discuss the relative merits of simulated maximum likelihood, Monte Carlo EM, Monte Carlo NewtonRaphson and stochastic approximation. Key words and phrases : Eciency, Monte Carlo EM, Monte Carlo NewtonRaphson, Rate of convergence, Simulated maximum likelihood, Stochastic approximation All three authors partially supported by NSF Grant DMS0072827. 1 1
Efficient estimation of transition rates between credit ratings from observations
"... at discrete time points ..."
Regressograms and meancovariance models for incomplete longitudinal data, American Statistician p. Revision Submitted
, 2011
"... Longitudinal studies are prevalent in biological and social sciences where subjects are measured repeatedly over time. Modeling the correlations and handling missing data are among the most challenging problems in analyzing such data. There are various methods for handling missing data, but databas ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
Longitudinal studies are prevalent in biological and social sciences where subjects are measured repeatedly over time. Modeling the correlations and handling missing data are among the most challenging problems in analyzing such data. There are various methods for handling missing data, but databased and graphical methods for modeling the covariance matrix of longitudinal data are relatively new. We adopt an approach based on the modified Cholesky decomposition of the covariance matrix which handles both challenges simultaneously. It amounts to formulating parametric models for the regression coefficients of the conditional mean and variance of each measurement given its predecessors. We demonstrate the roles of profile plots and regressograms in formulating joint meancovariance models for (in)complete longitudinal data. Applying these graphical tools to a casestudy of Fruit Fly Mortality data which has 22 % missing values reveals an “Sshape ” or logistic curve for the mean function and cubic polynomial models for the two factors of the modified Cholesky decomposition of the sample covariance matrix. A likelihoodbased method for estimating the parameters of the mean and covariance models using the EM algorithm is proposed. A simulation study and application to real data demonstrate that the estimation method works well for incomplete longitudinal data.
MUS81 Generates a Subset of MLH1MLH3–Independent Crossovers in Mammalian Meiosis
, 2008
"... Two eukaryotic pathways for processing doublestrand breaks (DSBs) as crossovers have been described, one dependent on the MutL homologs Mlh1 and Mlh3, and the other on the structurespecific endonuclease Mus81. Mammalian MUS81 has been implicated in maintenance of genomic stability in somatic cells ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
Two eukaryotic pathways for processing doublestrand breaks (DSBs) as crossovers have been described, one dependent on the MutL homologs Mlh1 and Mlh3, and the other on the structurespecific endonuclease Mus81. Mammalian MUS81 has been implicated in maintenance of genomic stability in somatic cells; however, little is known about its role during meiosis. Mus81deficient mice were originally reported as being viable and fertile, with normal meiotic progression; however, a more detailed examination of meiotic progression in Mus81null animals and WT controls reveals significant meiotic defects in the mutants. These include smaller testis size, a depletion of mature epididymal sperm, significantly upregulated accumulation of MLH1 on chromosomes from pachytene meiocytes in an interferenceindependent fashion, and a subset of meiotic DSBs that fail to be repaired. Interestingly, chiasmata numbers in spermatocytes from Mus81 2/2 animals are normal, suggesting additional integrated mechanisms controlling the two distinct crossover pathways. This study is the first indepth analysis of meiotic progression in Mus81nullizygous mice, and our results implicate the MUS81 pathway as a regulator of crossover
Standard Errors for EM Estimates in Generalized Linear Models with Random Effects
, 2000
"... A procedure is derived for computing standard errors of EM estimates in generalized linear models with random effects. Quadrature formulae are used to approximate the integrals in the EM algorithm, where two different approaches are pursued: GauHermite quadrature, in case of Gaussian random effects ..."
Abstract
 Add to MetaCart
A procedure is derived for computing standard errors of EM estimates in generalized linear models with random effects. Quadrature formulae are used to approximate the integrals in the EM algorithm, where two different approaches are pursued: GauHermite quadrature, in case of Gaussian random effects, and nonparametric maximum likelihood estimation for an unspecified random effect distribution. An approximation of the expected Fisher information matrix is derived from an expansion of the EM estimating equations. This allows for inferential arguments based on EM estimates, as demonstrated by an example and simulations. Keywords: EM algorithm, Estimating equations, GauHermite quadrature, Mixture model, Nonparametric maximum likelihood estimation, Random effect model. Institute of Statistics
A Hierarchical Statistical Model for Estimating Population Properties of Quantitative Genes
, 2002
"... Background: Earlier methods for detecting major genes responsible for a quantitative trait rely critically upon a wellstructured pedigree in which the segregation pattern of genes exactly follow Mendelian inheritance laws. However, for many outcrossing species, such pedigrees are not available a ..."
Abstract
 Add to MetaCart
Background: Earlier methods for detecting major genes responsible for a quantitative trait rely critically upon a wellstructured pedigree in which the segregation pattern of genes exactly follow Mendelian inheritance laws. However, for many outcrossing species, such pedigrees are not available and genes also display population properties.
The Keele Economics Department produces this series of research papers in order to stimulate discussion and invite feedback. Copyright remains with the authors.
"... Keywords J.E.L. Class Address Download 15 April 2003 We study the empirical relationship between the hiring rate, separation rate and starting wages. A practical empirical model is set up and estimated on Danish matched employeremployee longitudinal data for the period 19801995. We find (1) ..."
Abstract
 Add to MetaCart
Keywords J.E.L. Class Address Download 15 April 2003 We study the empirical relationship between the hiring rate, separation rate and starting wages. A practical empirical model is set up and estimated on Danish matched employeremployee longitudinal data for the period 19801995. We find (1) firm heterogeneity is important in all dimensions of our model: starting wages, time between hires and employment length (2) matching between firms and workers is exogenous (3) friction effects in the determination of wages are very small, which suggests that the Danish labour market is close to competitive.
Keele Economics Research Papers
"... The estimation of models involving discrete mixtures is a common practice in econometrics, for example to account for unobserved heterogeneity. However, the literature is relatively uninformative about the measurement of the precision of the parameters. This note provides an analytical expression fo ..."
Abstract
 Add to MetaCart
The estimation of models involving discrete mixtures is a common practice in econometrics, for example to account for unobserved heterogeneity. However, the literature is relatively uninformative about the measurement of the precision of the parameters. This note provides an analytical expression for the observed information matrix in terms of the gradient and hessian of the latent model when the number of components of the discrete mixture is known. This in turn allows for the estimation of the variance covariance matrix of the ML estimator of the parameters. I discuss further two possible applications of the result: the acceleration of the EM algorithm and the specification testing with the information matrix test.