Laplace Density Research Articles

Bayesian estimation for speech enhancement given a priori knowledge of clean speech phase

In this paper, STFT based speech enhancement algorithms based on estimation of short time spectral amplitudes are proposed. These algorithms use maximum likelihood, maximum a posterior and minimum mean square error (MMSE) estimators which respectively uses Laplace, Gamma and Exponential probability density functions as noise spectral amplitude priors and Nakagami distribution as speech spectral amplitude priors. The phase of noisy speech carries significant information to be retrieved and utilized. However, the undesired artifacts which are the resultant of the process do create many challenges. In this paper, the reconstructed phase is treated as an uncertain prior knowledge when deriving a joint MMSE estimate of the (C)omplex speech coefficients given (U)ncertain (P)hase information is proposed. The proposed phase reconstruction algorithm assists in generating a clean speech phase. The proposed estimator reduces undesired artifacts and also gives satisfactory values between noisy phase signal and estimate of prior phase and hence yields superior performance in the instrument measures, informal listening and speech quality.

A sandwich likelihood correction for Bayesian quantile regression based on the misspecified asymmetric Laplace density

A sandwich likelihood correction is proposed to remedy an inferential limitation of the Bayesian quantile regression approach based on the misspecified asymmetric Laplace density, by leveraging the benefits of the approach. Supporting theoretical results and simulations are presented.

Unsupervised learning via mixtures of skewed distributions with hypercube contours

Mixture models whose components have skewed hypercube contours are developed via a generalization of the multivariate shifted asymmetric Laplace density. Specifically, we develop mixtures of multiple scaled shifted asymmetric Laplace distributions. The component densities have a unique combination of features: they include a multivariate weight function and the marginal distributions are asymmetric Laplace. We use these mixtures of multiple scaled shifted asymmetric Laplace distributions for clustering applications, but they could be used in the supervised or semi-supervised paradigms. Parameter estimates are obtained via an expectation-maximization algorithm and the Bayesian information criterion is used for model selection. Simulated and real data sets are utilized to illustrate the approach and, in some cases, to visualize the skewed hypercube structure of the components.

Bayesian Estimation of Non-Gaussian Stochastic Volatility Models

In this paper, a general Non-Gaussian Stochastic Volatility model is proposed instead of the usual Gaussian model largely studied. We consider a new specification of SV model where the innovations of the return process have centered non-Gaussian error distribution rather than the standard Gaussian distribution usually employed. The model describes the behaviour of random time fluctuations in stock prices observed in the financial markets. It offers a response to better model the heavy tails and the abrupt changes observed in financial time series. We consider the Laplace density as a special case of non-Gaussian SV models to be applied to our data base. Markov Chain Monte Carlo technique, based on the bayesian analysis, has been employed to estimate the model’s parameters.

Posterior Consistency of Bayesian Quantile Regression Based on the Misspecified Asymmetric Laplace Density

We explore an asymptotic justification for the widely used and empirically verified approach of assuming an asymmetric Laplace distribution (ALD) for the response in Bayesian Quantile Regression. Based on empirical findings, Yu and Moyeed (2001) argued that the use of ALD is satisfactory even if it is not the true underlying distribution. We provide a justification to this claim by establishing posterior consistency and deriving the rate of convergence under the ALD misspecification. Related literature on misspecified models focuses mostly on i.i.d. models which in the regression context amounts to considering i.i.d. random covariates with i.i.d. errors. We study the behavior of the posterior for the misspecified ALD model with independent but non identically distributed response in the presence of non-random covariates. Exploiting the specific form of ALD helps us derive conditions that are more intuitive and easily seen to be satisfied by a wide range of potential true underlying probability distributions for the response. Through simulations, we demonstrate our result and also find that the robustness of the posterior that holds for ALD fails for a Gaussian formulation, thus providing further support for the use of ALD models in quantile regression.

Generalized double Pareto shrinkage

We propose a generalized double Pareto prior for Bayesian shrinkage estimation and inferences in linear models. The prior can be obtained via a scale mixture of Laplace or normal distributions, forming a bridge between the Laplace and Normal-Jeffreys' priors. While it has a spike at zero like the Laplace density, it also has a Student's t-like tail behavior. Bayesian computation is straightforward via a simple Gibbs sampling algorithm. We investigate the properties of the maximum a posteriori estimator, as sparse estimation plays an important role in many problems, reveal connections with some well-established regularization procedures, and show some asymptotic results. The performance of the prior is tested through simulations and an application.

Posterior Consistency of Bayesian Quantile Regression Under a Mis-Specified Likelihood Based on Asymmetric Laplace Density

We provide a theoretical justification for the widely used and yet only empirically verified approach of using Asymmetric Laplace Density(ALD) in Bayesian Quantile Regression. We derive sufficient conditions for posterior consistency of the quantile regression parameters even if the true underlying likelihood is not ALD, by considering both the case of random as well as non-random covariates. While existing literature on misspecified models address more general models, our approach of exploiting the specific form of ALD allows for a more direct derivation. We verify that the conditions so derived are satisfied by a wide range of potential true underlying probability distributions. We also show that posterior consistency holds even in the case of improper priors as long as the posterior is well defined. We demonstrate the working of the method using simulations.

Bayesian quantile regression for parametric nonlinear mixed effects models

We propose quantile regression (QR) in the Bayesian framework for a class of nonlinear mixed effects models with a known, parametric model form for longitudinal data. Estimation of the regression quantiles is based on a likelihood-based approach using the asymmetric Laplace density. Posterior computations are carried out via Gibbs sampling and the adaptive rejection Metropolis algorithm. To assess the performance of the Bayesian QR estimator, we compare it with the mean regression estimator using real and simulated data. Results show that the Bayesian QR estimator provides a fuller examination of the shape of the conditional distribution of the response variable. Our approach is proposed for parametric nonlinear mixed effects models, and therefore may not be generalized to models without a given model form.

Optimal estimation of $\ell_1$-regularization prior from a regularized empirical Bayesian risk standpoint

We address the problem of prior matrix estimation for the solution of$\ell_1$-regularized ill-posed inverse problems.From a Bayesian viewpoint, we show that such a matrix can be regarded as aninfluence matrix in a multivariate $\ell_1$-Laplace density function.Assuming a training set is given, the prior matrix design problem is cast as a maximum likelihood term with an additional sparsity-inducing term.This formulation results in an unconstrained yet nonconvex optimization problem. Memory requirements as well as computation of the nonlinear,nonsmooth sub-gradient equations are prohibitive for large-scale problems. Thus, we introduce an iterative algorithm todesign efficient priors for such large problems. We further demonstrate that the solutions of ill-posed inverse problems by incorporation of $\ell_1$-regularization using the learned prior matrix perform generally better than commonly used regularization techniques where the prior matrix is chosen a-priori.

On Construction and Simulation of Autoregressive Sources With Near-Laplace Marginals

In this paper, we focus upon the problem of modeling and simulation of stationary non-Gaussian time series. In particular, we consider a first order autoregressive process whose marginal distribution is close to the Laplace density. This model allows us to simulate correlated non-Gaussian signals typically appearing in speech analysis, compression, and noise synthesis. The Monte Carlo rejection method is applied to develop efficient algorithms for simulation of the proposed autoregressive process. We also extend our theory and algorithms to the related issue of constructing a correlated bivariate time-series model with near-Laplace margins. A theoretical analysis of the average complexity of the proposed simulation algorithms is included.

Inversion of Analytic Characteristic Functions and Infinite Convolutions of Exponential and Laplace Densities

We prove that certain quotients of entire functions are characteristic functions. Under some conditions, the probability measure corresponding to a characteristic function of that type has a density which can be expressed as a generalized Dirichlet series, which in turn is an infinite linear combination of exponential or Laplace densities. These results are applied to several examples.

Seminormal composition operators on [formula omitted] spaces induced by matrices: The Laplace density case

Bounded composition operators with matrix symbols on L 2 ( μ ) , where μ is a positive Borel measure with a Laplace type density function on the d-dimensional Euclidean space, are studied. Hyponormality and subnormality of adjoints of such operators are characterized explicitly in terms of their symbols.

Long-tail longitudinal modeling of insurance company expenses

The insurance industry is known to have high operating expenses in the financial services sector. Insurers, investors and regulators are interested in models to understand the behavior of expenses. However, the current practice ignores skewness, occasional negative values as well as their temporal dependence. Addressing these three features, this paper develops a longitudinal model of insurance company expenses that can be used for prediction, to identify unusual behavior, and to measure firm efficiency. Specifically, we use a three-parameter asymmetric Laplace density for the marginal distribution of insurers’ expenses in each year. Copula functions are employed to accommodate their temporal dependence. As a function of explanatory variables, the location parameter allows us to analyze an insurer’s expenses in light of the firm’s characteristics. Our model can be interpreted as a longitudinal quantile regression. The analysis is performed using property–casualty insurance company data from the National Association of Insurance Commissioners of years 2001–2006. Due to the long-tailed nature of insurers’ expenses, two alternative approaches are proposed to improve the performance of the longitudinal quantile regression model: rescaling and transformation. Predictive densities are derived that allow one to compare the predictions for individual insurers in a hold-out-sample. Both predictive models are shown to be reasonable with the rescaling method outperforming the transformation method. Compared with standard longitudinal models, our model is shown to be superior in identifying insurers’ unusual behavior.

An Alternative to the Bessel K form Distribution

We study an alternative to the Bessel K form (BKF) distribution that results when instead of gamma, the scale random variable in the Gaussian scale mixture (GSM) parameterization follows a generalized exponential distribution. The new distribution is expressed as a sum of simpler Laplace densities. Some properties of this distribution are provided. We investigate parameter estimation under the additive Gaussian noise model using simple cascades of estimators. Exact and approximate Bayesian estimators are derived for the generalized K form (GKF) in the spherically contoured case that make use of the generalized incomplete Gamma function.

Modeling of Sunspot Numbers by a Modified Binary Mixture of Laplace Distribution Functions

This paper presents a new approach for describing the shape of 11-year sunspot cycles by considering the monthly averaged values. This paper also brings out a prediction model based on the analysis of 22 sunspot cycles from the year 1749 onward. It is found that the shape of the sunspot cycles with monthly averaged values can be described by a functional form of modified binary mixture of Laplace density functions, modified suitably by introducing two additional parameters in the standard functional form. The six parameters, namely two locations, two scales, and two area parameters, characterize this model. The nature of the estimated parameters for the sunspot cycles from 1749 onward has been analyzed and finally we arrived at a sufficient set of the parameters for the proposed model. It is seen that this model picks up the sunspot peaks more closely than any other model without losing the match at other places at the same time. The goodness of fit for the proposed model is also computed with the Hathaway – Wilson – Reichmann \(\overline{\chi}\) measure, which shows, on average, that the fitted model passes within 0.47 standard deviations of the actual averaged monthly sunspot numbers.

Bayesian Estimation of Bessel K Form Random Vectors in AWGN

We present new Bayesian estimators for spherically-contoured Bessel K form (BKF) random vectors in additive white Gaussian noise (AWGN). The derivations are an extension of existing results for the scalar BKF and multivariate Laplace (MLAP) densities. MAP and MMSE estimators are derived. We show that the MMSE estimator can be written in exact form in terms of the generalized incomplete Gamma function. Computationally efficient approximations are given. We compare the proposed exact and approximate MMSE estimators with recent results using the BKF density, both in terms of the shrinkage rules and the associated mean-square error.

Quantile regression for longitudinal data using the asymmetric Laplace distribution

In longitudinal studies, measurements of the same individuals are taken repeatedly through time. Often, the primary goal is to characterize the change in response over time and the factors that influence change. Factors can affect not only the location but also more generally the shape of the distribution of the response over time. To make inference about the shape of a population distribution, the widely popular mixed-effects regression, for example, would be inadequate, if the distribution is not approximately Gaussian. We propose a novel linear model for quantile regression (QR) that includes random effects in order to account for the dependence between serial observations on the same subject. The notion of QR is synonymous with robust analysis of the conditional distribution of the response variable. We present a likelihood-based approach to the estimation of the regression quantiles that uses the asymmetric Laplace density. In a simulation study, the proposed method had an advantage in terms of mean squared error of the QR estimator, when compared with the approach that considers penalized fixed effects. Following our strategy, a nearly optimal degree of shrinkage of the individual effects is automatically selected by the data and their likelihood. Also, our model appears to be a robust alternative to the mean regression with random effects when the location parameter of the conditional distribution of the response is of interest. We apply our model to a real data set which consists of self-reported amount of labor pain measurements taken on women repeatedly over time, whose distribution is characterized by skewness, and the significance of the parameters is evaluated by the likelihood ratio statistic.

Robust minimax detection of a weak signal in noise with a bounded variance and density value at the center of symmetry

In practical communication environments, it is frequently observed that the underlying noise distribution is not Gaussian and may vary in a wide range from short-tailed to heavy-tailed forms. To describe partially known noise distribution densities, a distribution class characterized by the upper-bounds upon a noise variance and a density dispersion in the central part is used. The results on the minimax variance estimation in the Huber sense are applied to the problem of asymptotically minimax detection of a weak signal. The least favorable density minimizing Fisher information over this class is called the Weber-Hermite density and it has the Gaussian and Laplace densities as limiting cases. The subsequent minimax detector has the following form: i) with relatively small variances, it is the minimum L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -norm distance rule; ii) with relatively large variances, it is the L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1 </sub> -norm distance rule; iii) it is a compromise between these extremes with relatively moderate variances. It is shown that the proposed minimax detector is robust and close to Huber's for heavy-tailed distributions and more efficient than Huber's for short-tailed ones both in asymptotics and on finite samples

Do liquidity constraints matter in explaining firm size and growth? Some evidence from the Italian manufacturing industry

The article investigates whether liquidity constraints affect firm size and growth dynamics of Italian manufacturing firms. Panel-data regressions and distribution analyses show that (i) liquidity constraints engender a negative effect on growth once one controls for size; (ii) smaller firms grow more after controlling for liquidity constraints; and (iii) the stronger liquidity constraints, the more size negatively affects firm growth. Furthermore, we find that financial constraints help in better explaining the relationship between firm growth and age, conditional on size. Finally, our data indicate that size distributions depart from log-normality, and growth rates are well approximated by Laplace densities.

EbayesThresh:RPrograms for Empirical Bayes Thresholding

Suppose that a sequence of unknown parameters is observed sub ject to independent Gaussian noise. The EbayesThresh package in the S language implements a class of Empirical Bayes thresholding methods that can take advantage of possible sparsity in the sequence, to improve the quality of estimation. The prior for each parameter in the sequence is a mixture of an atom of probability at zero and a heavy-tailed density. Within the package, this can be either a Laplace (double exponential) density or else a mixture of normal distributions with tail behavior similar to the Cauchy distribution. The mixing weight, or sparsity parameter, is chosen automatically by marginal maximum likelihood. If estimation is carried out using the posterior median, this is a random thresholding procedure; the estimation can also be carried out using other thresholding rules with the same threshold, and the package provides the posterior mean, and hard and soft thresholding, as additional options. This paper reviews the method, and gives details (far beyond those previously published) of the calculations needed for implementing the procedures. It explains and motivates both the general methodology, and the use of the EbayesThresh package, through simulated and real data examples. When estimating the wavelet transform of an unknown function, it is appropriate to apply the method level by level to the transform of the observed data. The package can carry out these calculations for wavelet transforms obtained using various packages in R and S-PLUS. Details, including a motivating example, are presented, and the application of the method to image estimation is also explored. The final topic considered is the estimation of a single sequence that may become progressively sparser along the sequence. An iterated least squares isotone regression method allows for the choice of a threshold that depends monotonically on the order in which the observations are made. An alternative possibility, also discussed in detail, is a particular parametric dependence of the sparsity parameter on the position in the sequence.