Seemingly Unrelated Ridge Regression in Semiparametric Models

This article is concerned with the problem of multicollinearity in the linear part of a seemingly unrelated semiparametric (SUS) model. It is also suspected that some additional non stochastic linear constraints hold on the whole parameter space. In the sequel, we propose semiparametric ridge and non ridge type estimators combining the restricted least squares methods in the model under study. For practical aspects, it is assumed that the covariance matrix of error terms is unknown and thus feasible estimators are proposed and their asymptotic distributional properties are derived. Also, necessary and sufficient conditions for the superiority of the ridge-type estimator over the non ridge type estimator for selecting the ridge parameter K are derived. Lastly, a Monte Carlo simulation study is conducted to estimate the parametric and nonparametric parts. In this regard, kernel smoothing and cross validation methods for estimating the nonparametric function are used.     

Arbitrage Pricing Theory as a Restricted Nonlinear Multivariate Regression Model Iterated Nonlinear Seemingly Unrelated Regression Estimates

By replacing the unknown random factors of factor analysis with observed macroeconomic variables, the arbitrage pricing theory (APT) is recast as a multivariate nonlinear regression model with across-equation restrictions. An explicit theoretical justification for the inclusion of an arbitrary, well-diversified market index is given. Using monthly returns on 70 stocks, iterated nonlinear seemingly unrelated regression techniques are employed to obtain joint estimates of asset sensitivities and their associated APT risk “prices.” Without the assumption of normally distributed errors, these estimators are strongly consistent and asymptotically normal. With the additional assumption of normal errors, they are also full-information maximum likelihood estimators. Classical asymptotic nonlinear nested hypothesis tests are supportive of the APT with measured macroeconomic factors.     

Some Asymptotic Properties of Ridge Regression in a System of Seemingly Unrelated Regression Equations

A discussion is made of asymptotic properties of an Operational Ordinary Ridge Regression estimator and comparison is made with the Operational Generalized Least Squares estimator. Also, some simulation experiments are carried showing efficiency gains can be made through the use of de Ridge estimator.     

A Comparison of REML and Covariance Adjustment Method in the Estimation of Growth Curve Models

The classical growth curve model is considered when one continuous characteristic is measured at q time points. The covariance adjusted estimator of growth curve parameters is the OLS estimator adjusted using analysis of covariance. The covariates are obtained from functions of within individuals error contrasts. On the other hand, REML estimators emerge from maximization of the likelihood of OLS residuals. We compare the efficiency of estimators of growth curve parameters obtained by REML with that of covariance-adjusted least squares estimators with covariates selected via CAIC.     

Efficient Estimation of the Seemingly Unrelated Regression Cointegration Model and Testing for Purchasing Power Parity

This paper studies the efficient estimation of seemingly unrelated linear models with integrated regressors and stationary errors. We consider two cases. The first one has no common regressor among the equations. In this case, we show that by adding leads and lags of the first differences of the regressors and estimating this augmented dynamic regression model by generalized least squares using the long-run covariance matrix, we obtain an efficient estimator of the cointegrating vector that has a limiting mixed normal distribution. In the second case we consider, there is a common regressor to all equations, and we discuss efficient minimum distance estimation in this context. Simulation results suggests that our new estimator compares favorably with others already proposed in the literature. We apply these new estimators to the testing of the proportionality and symmetry conditions implied by purchasing power parity (PPP) among the G-7 countries. The tests based on the efficient estimates easily reject the joint hypotheses of proportionality and symmetry for all countries with either the United States or Germany as numeraire. Based on individual tests, our results suggest that Canada and Germany are the most likely countries for which the proportionality condition holds, and that Italy and Japan for the symmetry condition relative to the United States.     

Efficient Estimation of the Seemingly Unrelated Regression Cointegration Model and Testing for Purchasing Power Parity

Abstract

This paper studies the efficient estimation of seemingly unrelated linear models with integrated regressors and stationary errors. We consider two cases. The first one has no common regressor among the equations. In this case, we show that by adding leads and lags of the first differences of the regressors and estimating this augmented dynamic regression model by generalized least squares using the long-run covariance matrix, we obtain an efficient estimator of the cointegrating vector that has a limiting mixed normal distribution. In the second case we consider, there is a common regressor to all equations, and we discuss efficient minimum distance estimation in this context. Simulation results suggests that our new estimator compares favorably with others already proposed in the literature. We apply these new estimators to the testing of the proportionality and symmetry conditions implied by purchasing power parity (PPP) among the G-7 countries. The tests based on the efficient estimates easily reject the joint hypotheses of proportionality and symmetry for all countries with either the United States or Germany as numeraire. Based on individual tests, our results suggest that Canada and Germany are the most likely countries for which the proportionality condition holds, and that Italy and Japan for the symmetry condition relative to the United States.     

Equivalence of maximum likelihood estimation and iterative two-stage estimation for seemingly unrelated regression models

The seemingly unrelated regression model is viewed in the context of repeated measures analysis. Regression parameters and the variance-covariance matrix of the seemingly unrelated regression model can be estimated by using two-stage Aitken estimation. The first stage is to obtain a consistent estimator of the variance-covariance matrix. The second stage uses this matrix to obtain the generalized least squares estimators of the regression parameters. The maximum likelihood (ML) estimators of the regression parameters can be obtained by performing the two-stage estimation iteratively. The iterative two-stage estimation procedure is shown to be equivalent to the EM algorithm (Dempster, Laird, and Rubin, 1977) proposed by Jennrich and Schluchter (1986) and Laird, Lange, and Stram (1987) for repeated measures data. The equivalence of the iterative two-stage estimator and the ML estimator has been previously demonstrated empirically in a Monte Carlo study by Kmenta and Gilbert (1968). It does not appear to be widely known that the two estimators are equivalent theoretically. This paper demonstrates this equivalence.     

Nonparametric Growth Curve Model with Local Linear Approximation

The growth curve model has been developed for longitudinal data, and its time trend is usually described by polynomials. However, it is difficult to interpret each coefficient of the polynomials with higher degrees, even when the number of repetitions is sufficiently large. We propose herein an alternative growth curve model having time-varying coefficients.     

Nonlinear Quantile Regression Estimation of Longitudinal Data

This article examines a weighted version of the quantile regression estimator as defined by Koenker and Bassett (1978 Koenker , R. , Bassett , G. ( 1978 ). Regression quantiles . Econometrica 46 : 33 – 50 .[Crossref], [Web of Science ®] , [Google Scholar]), adjusted to the case of nonlinear longitudinal data. Using a four-parameter logistic growth function and error terms following an AR(1) model, different weights are used and compared in a simulation study. The findings indicate that the nonlinear quantile regression estimator is performing well, especially for the median regression case, that the differences between the weights are small, and that the estimator performs better when the correlation in the AR(1) model increases. A comparison is also made with the corresponding mean regression estimator, which is found to be less robust. Finally, the estimator is applied to a data set with growth patterns of two genotypes of soybean, which gives some insights into how the quantile regressions provide a more complete picture of the data than the mean regression.     

A Simulation Comparison of Approximate Tests for Fixed Effects in Random Coefficients Growth Curve Models

Often, the response variables on sampling units are observed repeatedly over time. The sampling units may come from different populations, such as treatment groups. This setting is routinely modeled by a random coefficients growth curve model, and the techniques of general linear mixed models are applied to address the primary research aim. An alternative approach is to reduce each subject’s data to summary measures, such as within-subject averages or regression coefficients. One may then test for equality of means of the summary measures (or functions of them) among treatment groups. Here, we compare by simulation the performance characteristics of three approximate tests based on summary measures and one based on the full data, focusing mainly on accuracy of p-values. We find that performances of these procedures can be quite different for small samples in several different configurations of parameter values. The summary-measures approach performed at least as well as the full-data mixed models approach.     

The Stabilisation of Model Parameter Estimates from Repeated Surveys with Rare Observations

In many surveys, the domains of study are small and the samples that carry information on a domain can be very small indeed. If the survey is conducted repeatedly there is often a high degree of overlap in samples over time. We show how to use the richness of information over time to compensate for the paucity of cross‐sectional information. We propose a model‐based estimator of the population total which makes use of stabilised parameter estimates that combine information from different survey periods that are adjacent in time. The motivating example for this research was the ProdCom survey as implemented in the UK.     

General class of covariance structures for two or more repeated factors in longitudinal data analysis

The main difficulty in parametric analysis of longitudinal data lies in specifying covariance structure. Several covariance structures, which usually reflect one series of measurements collected over time, have been presented in the literature. However there is a lack of literature on covariance structures designed for repeated measures specified by more than one repeated factor. In this paper a new, general method of modelling covariance structure based on the Kronecker product of underlying factor specific covariance profiles is presented. The method has an attractive interpretation in terms of independent factor specific contribution to overall within subject covariance structure and can be easily adapted to standard software.     

The Mean Squared Error Optimum Design Criterion for Parameter Estimation in Nonlinear Regression Models

In the context of nonlinear regression models, we propose an optimal experimental design criterion for estimating the parameters that account for the intrinsic and parameter-effects nonlinearity. The optimal design criterion proposed in this article minimizes the determinant of the mean squared error matrix of the parameter estimator that is quadratically approximated using the curvature array. The design criterion reduces to the D-optimal design criterion if there are no intrinsic and parameter-effects nonlinearity in the model, and depends on the scale parameter estimator and on the reparameterization used. Some examples, using a well known nonlinear kinetics model, demonstrate the application of the proposed criterion to nonsequential design of experiments as compared with the D-optimal criterion.     

正交多项式回归模型参数估计的格子设计及其最小二乘估计

对二元二次多项式回归模型进行预先正交化处理,提出使用因子空间的N等份的格子设计的观点,推导出了信息矩阵的一般性结构,并给出了该设计的非退化条件及其最小二乘估计和对应的协方差矩阵。     

Marginal Correlation from Logit- and Probit-Beta-Normal Models for Hierarchical Binary Data

In hierarchical data settings, be it of a longitudinal, spatial, multi-level, clustered, or otherwise repeated nature, often the association between repeated measurements attracts at least part of the scientific interest. Quantifying the association frequently takes the form of a correlation function, including but not limited to intraclass correlation. Vangeneugden et al. (2010 Vangeneugden, T., Molenberghs, G., Laenen, A., Geys, H., Beunckens, C., Sotto, C. (2010). Marginal correlation in longitudinal binary data based on generalized linear mixed models. Communi. Stati. Theory &; Methods. 39:3542–3557. [Google Scholar]) derived approximate correlation functions for longitudinal sequences of general data type, Gaussian and non-Gaussian, based on generalized linear mixed-effects models. Here, we consider the extended model family proposed by Molenberghs et al. (2010 Molenberghs, G., Verbeke, G., Demétrio, C., Vieira, A. (2010). A family of generalized linear models for repeated measures with normal and conjugate random effects. Stat. Sci. 25:325–347.[Crossref], [Web of Science ®] , [Google Scholar]). This family flexibly accommodates data hierarchies, intra-sequence correlation, and overdispersion. The family allows for closed-form means, variance functions, and correlation function, for a variety of outcome types and link functions. Unfortunately, for binary data with logit link, closed forms cannot be obtained. This is in contrast with the probit link, for which such closed forms can be derived. It is therefore that we concentrate on the probit case. It is of interest, not only in its own right, but also as an instrument to approximate the logit case, thanks to the well-known probit-logit ‘conversion.’ Next to the general situation, some important special cases such as exchangeable clustered outcomes receive attention because they produce insightful expressions. The closed-form expressions are contrasted with the generic approximate expressions of Vangeneugden et al. (2010 Vangeneugden, T., Molenberghs, G., Laenen, A., Geys, H., Beunckens, C., Sotto, C. (2010). Marginal correlation in longitudinal binary data based on generalized linear mixed models. Communi. Stati. Theory &; Methods. 39:3542–3557. [Google Scholar]) and with approximations derived for the so-called logistic-beta-normal combined model. A simulation study explores performance of the method proposed. Data from a schizophrenia trial are analyzed and correlation functions derived.     

Power and Sample Size for Fixed-Effects Inference in Reversible Linear Mixed Models

ABSTRACT

Despite the popularity of the general linear mixed model for data analysis, power and sample size methods and software are not generally available for commonly used test statistics and reference distributions. Statisticians resort to simulations with homegrown and uncertified programs or rough approximations which are misaligned with the data analysis. For a wide range of designs with longitudinal and clustering features, we provide accurate power and sample size approximations for inference about fixed effects in the linear models we call reversible. We show that under widely applicable conditions, the general linear mixed-model Wald test has noncentral distributions equivalent to well-studied multivariate tests. In turn, exact and approximate power and sample size results for the multivariate Hotelling–Lawley test provide exact and approximate power and sample size results for the mixed-model Wald test. The calculations are easily computed with a free, open-source product that requires only a web browser to use. Commercial software can be used for a smaller range of reversible models. Simple approximations allow accounting for modest amounts of missing data. A real-world example illustrates the methods. Sample size results are presented for a multicenter study on pregnancy. The proposed study, an extension of a funded project, has clustering within clinic. Exchangeability among the participants allows averaging across them to remove the clustering structure. The resulting simplified design is a single-level longitudinal study. Multivariate methods for power provide an approximate sample size. All proofs and inputs for the example are in the supplementary materials (available online).     

空间回归模型估计中的最小二乘法

空间回归模型由于引入了空间地理信息而使得其参数估计变得复杂,因为主要采用最大似然法,致使一般人认为在空间回归模型参数估计中不存在最小二乘法。通过分析空间回归模型的参数估计技术,研究发现,最小二乘法和最大似然法分别用于估计空间回归模型的不同的参数,只有将两者结合起来才能快速有效地完成全部的参数估计。数理论证结果表明,空间回归模型参数最小二乘估计量是最佳线性无偏估计量。空间回归模型的回归参数可以在估计量为正态性的条件下而实施显著性检验,而空间效应参数则不可以用此方法进行检验。     

Approximation of the Distribution of the Location Parameter in the Growth Curve Model

Abstract.  In this paper an Edgeworth-type approximation of order O(n ⁻² ) to the density of the estimator of the location parameter in the growth curve model has been derived. The approximation is a mixture of a normal and a Kotz-type distribution, thus being an elliptical distribution. A condition for unimodality of the mixture was found and marginal distribution of a subvector of the mixture distribution was derived. Finally, a small example was given to demonstrate an application of the approximation.     

The Empirical Likelihood Inference of a Regression Parameter in Censored Partial Linear Models Based on a Piecewise Polynomial

This article aims at making an empirical likelihood inference of regression parameter in partial linear model when the response variable is right censored randomly. The present studies are mainly designed to use empirical likelihood (EL) method based on synthetic dependent data, and the result cannot be applied directly due to the unknown weights in it. In this paper, we introduce a censored empirical log-likelihood ratio and demonstrate that its limiting distribution is a standard chi-square distribution. The estimating procedure of β is developed based on piecewise polynomial method. As a result, the p-value of test and the confidence interval can be obtained without estimating other quantities. Some simulation studies are conducted to highlight the performance of the proposed EL method, and the results show a good performance. Finally, we apply our method into the real example of multiple myeloma data and show the proof of theorem.     

A Note on Criterion-Robust Optimal Designs for Model Discrimination and Parameter Estimation in Polynomial Regression Models

Consider the problem of discriminating between the polynomial regression models on [?1, 1] and estimating parameters in the models. Zen and Tsai (2002 Zen , M. M. , Tsai , M. H. ( 2002 ). Some criterion-robust optimal designs for the dual problem of model discrimination and parameter estimation . Sankhya Ind. J. Statist. 64 : (Series B, Pt. 3) : 322 – 338 . [Google Scholar]) proposed a multiple-objective optimality criterion, M _γ-criterion, which uses weight γ (0 ≤ γ ≤ 1) for model discrimination and α = β = (1 ? γ)/2 for parameter estimation in each model. In this article, we generalize it to a wider setup with different values of α and β. For instance, α = 2 β suggests that the “smaller” model is more likely to be the true model. Using similar techniques, the corresponding criterion-robust optimal design is investigated. A study for the original criterion-robust optimal design with α = β, through M-efficiency, shows that it is good enough for any wider setup.