The cross-covariance matrix of observation vectors in two linear statistical models need not be zero matrix. In such a case the problem is to find explicit expressions for the best linear unbiased estimators of both model parameters and estimators of variance components in the simplest structure of the covariance matrix. Univariate and multivariate forms of linear models are dealt with.
The paper considers the problem of consistent variable selection in parametic models with the use of stepdown multiple hypothesis procedures. Our approach completes the results of Bunea et al. [J. Statist. Plann. Inference 136 (2006)]. A simulation study supports the results obtained.
In this work, we consider the general linear model or its variants with the ordinary least squares, generalised least squares or restricted least squares estimators of the regression coefficients and variance. We propose a newly unified set of definitions for local sensitivity for both situations, one for the estimators of the regression coefficients, and the other for the estimators of the variance. Based on these definitions, we present the estimators’ sensitivity results.We include brief remarks...
The singular mixed linear model with constraints is investigated with respect to an influence of inaccurate variance components on a decrease of the confidence level. The algorithm for a determination of the boundary of the insensitivity region is given. It is a set of all shifts of variance components values which make the tolerated decrease of the confidence level only. The problem about geometrical characterization of the confidence domain is also presented.
We propose a sequential monitoring scheme for detecting a change in scale. We consider a stable historical period of length . The goal is to propose a test with asymptotically small probability of false alarm and power 1 as the length of the historical period tends to infinity. The asymptotic distribution under the null hypothesis and consistency under the alternative hypothesis is derived. A small simulation study illustrates the finite sample performance of the monitoring scheme.
In this paper sign and Wilcoxon tests for testing the null hypothesis of quadratic regression versus the alternative, cubic regression are proposed. It is shown that in the case of a simple design consisting of multiple Y observations at each of the four levels of x, the proposed tests perform reasonably well as compared to their parametric competitors, while in the case of a general design consisting of a large number of levels of x, the loss in Pitman efficiency is considerable. However their...
Este trabajo presenta un procedimiento para hacer robusto el algoritmo recursivo de Plackett-Kalman para el modelo lineal, incorporándole medidas diagnósticas que indiquen la influencia potencial y real de cada nueva observación en los parámetros del modelo. Se describe cómo calcular recursivamente el estadístico D2 de Cook, la distancia de Mahalanobis de cada nueva observación al centro de gravedad de la ya incluidas, y un contraste, basado en los residuos recursivos, de que la nueva observación...
Regression- and scale-invariant -test procedures for detection of structural changes in linear regression model was developed and their limit behavior under the null hypothesis was studied in Hušková [9]. In the present paper the limit behavior under local alternatives is studied. More precisely, it is shown that under local alternatives the considered test statistics have asymptotically normal distribution.
Regression and scale invariant -test procedures are developed for detection of structural changes in linear regression model. Their limit properties are studied under the null hypothesis.
In this paper, we consider a comparison problem of predictors in the context of linear mixed models. In particular, we assume a set of different seemingly unrelated linear mixed models (SULMMs) allowing correlations among random vectors across the models. Our aim is to establish a variety of equalities and inequalities for comparing covariance matrices of the best linear unbiased predictors (BLUPs) of joint unknown vectors under SULMMs and their combined model. We use the matrix rank and inertia...
The paper deals with the experimental design which is optimal in the following sense: it satisfies the cost requirements simultaneously with a satisfactory precision of estimates. The underlying regression model is quadratic. The estimates of unknown parameters of the model are explicitly derived.
The paper discusses applications of permutation arguments in testing problems in linear models. Particular attention will be paid to the application in L₁-test procedures. Theoretical results will beaccompanied by a simulation study.
Some remarks to problems of point and interval estimation, testing and problems of outliers are presented in the case of multivariate regression model.