-
作者：HALL, P; WAND, MP
摘要：We propose a technique for nonparametric discrimination in which smoothing parameters are chosen jointly, according to a criterion based on the difference between two densities. The approach is suitable for categorical, continuous and mixed data, and uses information from both populations to determine the smoothing parameter for any one population. In the case of categorical data, optimal performance is sometimes achieved using negative smoothing parameters, a property which does not emerge if...
-
作者：ZHENG, ZK
摘要：A sequential plan with censored data for nonparametric testing of a hypothesis H0: F(t) = F0(t) against H1: F(t) .ltoreq. F0(t), where the inequality holds strictly for some t, is discussed. A large-sample theorem due to Gill (1983) is used and a Kolmogorov-type test is established.
-
作者：HOUGAARD, P
-
作者：PHILLIPS, PCB; PERRON, P
作者单位：Universite de Montreal
摘要：This paper proposes new tests for detecting the presence of a unit root in quite general time series models. Our approach is nonparametric with respect to nuisance parameters and thereby allows for a very wide class of weakly dependent and possibly heterogeneously distributed data. The tests accommodate models with a fitted drift and a time trend so that they may be used to discriminate beween unit root nonstationarity and stationarity about a deterministic trend. The limiting distributions of...
-
作者：GAIL, MH; TAN, WY; PIANTADOSI, S
作者单位：National Institutes of Health (NIH) - USA; NIH National Cancer Institute (NCI); University of Memphis
摘要：We propose a test of the null hypothesis of no treatment effect in a randomized clinical trial that is based on the randomization distribution of residuals. These residuals result from regressing the response on covariates, but not treatment. In contrast to model-based score tests, this procedure maintains nominal size when the model is misspecified, and, in particular, when relevant covariates are omitted from the regression. The efficiency of the procedure is evaluated for regressions with s...
-
作者：CHALONER, K; BRANT, R
作者单位：University of Toronto
摘要：An approach to detecting outliers in a linear model is developed. An outer outlier is defined to be an observation with a large random error, generated by the linear model under consideration. Outliers are detected by examining the posterior distribution of the random errors. An augmented residual plot is also suggested as a graphical aid in finding outliers.
-
作者：CONNOLLY, MA; LIANG, KY
摘要：A class of conditional logistic regression models for clustered binary data is considered. This includes the polychotomous logistic model of Rosner (1984) as a special case. Properties such as the joint distribution and pairwise odds ratio are investigated. A class of easily computed estimating functions is introduced which is shown to have high efficiency compared to the computationally intensive maximum likelihood approach. An example on chronic obstructive pulmonary disease among sibs is pr...
-
作者：GLASBEY, CA
摘要：For parameter estimators in linear models, variance estimates are proposed which are positive-semidefinite quadratic forms like the conventional ones but are less dependent on assumptions about error variance. Examples are given of their use in spatial analyses of field trials and analyses of series of trials.
-
作者：BRECKLING, J; CHAMBERS, R
摘要：It is well known that a M-estimator of the centre of symmetry .theta. of a symmetric distribution can be defined in terms of either a continuous symmetric loss function .rho. or the associated influence function .psi.. This estimator is robust if .psi. is bounded. In this paper, we develop a generalization of the M-concept to estimation of a quantile analogue of .theta. called a M-quantile, by introducing a particular kind of asymmetry into .psi.. A natural consequence is that the M-quantile p...
-
作者：BARNDORFFNIELSEN, OE
作者单位：Australian National University
摘要：It is well known that Bartlett adjustment reduces level-error of the likelihood ratio statistic from order n-1 to order n-3/2. In the present note we show that level-error of the adjusted statistic is actually order n-2.
