-
作者：PETTIT, LI; YOUNG, KDS
作者单位：University of Surrey
摘要：In this paper we consider a measure of the effect of single observations on a logarithmic Bayes factor defined via the difference in the logarithms of the Bayes factors conditional first on all the data and then omitting an observation. The measure is related to the conditional predictive ordinate. The form of the measure and examples of its use are presented for a variety of situations, normal samples, linear linear models, log linear models and the checking of distributional assumptions.
-
作者：FU, YX; CURNOW, RN
摘要：Maximum likelihood estimation of the locations of changes in sequences of independent categorical random variables when there are known lower bounds on the lengths of the sub-sequences between the change points in discussed. A method is developed which finds one of the maximum likelihood solutions. The problem of estimating the number of changed segments is discussed and some numerical results about the precision of the estimates of the locations of changes are presented. The method also allow...
-
作者：SHARPLES, LD
摘要：By their nature, hierarchical models involve a considerable number of structural and distributional assumptions, for example, normality and linearity are common assumptions. Considerations of robustness therefore suggest that some form of model elaboration is necessary. However, it is not clear how to achieve this in the already complex setting of hierarchial models whilst maintaining reasonable tractability in the analysis. We provide a comprehensive account of the way in which methods based ...
-
作者：BROWN, EN
摘要：The harmonic regression signal plus noise model is used extensively in applied statistics. This note establishes the joint asymptotic normality of the generalized least squares estimates of the harmonic regression parameters and gives an explicit expression for the limiting covariance matrix. It also clarifies misstatements concerning the model''s properties that have appeared in the literature.
-
作者：HORN, PS
摘要：In this study we propose quantile estimators for a variety of skewed distributions. The proposed estimators are based on the biweight estimators for symmetric populations (Horn, 1988). These achieve good coverage across a wide variety of skewed populations, even for sample sizes as small as ten.
-
作者：BEGG, CB
摘要：Wei (1988) analyzed data from a clinical trial in which an urn-sampling model was used to allocate patients to treatments. The trial resulted in 11 patients being allocated to the experimental treatment, all successes, and with one patient allocated to the control treatment, a failure. Wei analyzed these data using a randomization algorithm and concluded that the results were almost significant at p = 0.051. He asserted that an analysis which ignored the design and presumed complete randomizat...
-
作者：HELLER, G; SIMONOFF, JS
作者单位：New York University
摘要：Several methods of estimating parameters in a linear regression when the response variable is censored are compared. The estimator of Buckley and James (1979) is preferred. Further, it is shown that the strength of the regression, the support and distribution of the censoring variable, and the magnitude of the slope all have an effect on the relative performance of the estimators.
-
作者：ZHAO, LP; PRENTICE, RL
摘要：A model of quadratic exponential form is parameterized in terms of marginal means and pairwise correlations for the regression analysis of correlated binary data. Pseudo-maximum likelihood methods using a special case termed the multiplicative model are proposed, but are noted to be computationally unattractive if the ''blocks'' of correlated responses are at all large. On the other hand score estimating functions for mean and correlation parameters are shown to be of a particularly simple for...
-
作者：SCHAID, DJ; WIEAND, S; THERNEAU, TM
摘要：A two-stage design is presented which enables the screening, at the first stage, of several new experimental treatments for survival improvement over a standard regimen. Only promising treatments are carried forward to the second stage for definitive evaluation. Our procedure is to minimize the number of patients expected to be accrued to new regimens which do not offer a survival benefit over the standard regimen, subject to constraints of specified alpha-error, power, and a stopping rule whi...
-
作者：ROTNITZKY, A; JEWELL, NP
作者单位：University of California System; University of California Berkeley; University of California System; University of California Berkeley
摘要：Generalized and ''working''Wald and score tests for regression coefficients in the class of semiparametric marginal generalized linear models for cluster correlated data (Liang and Zeger, 1986) are proposed, and their asymptotic distribution examined. In addition, the asymptotic distribution of the naive likelihood ratio test, or deviance difference, is presented. Following Rao and Scott (1984), we propose simple adjustments to such ''working'' tests. The asymptotic distributions of the ''work...
