-
作者：BARTMANN, FC; BLOOMFIELD, P
作者单位：Princeton University
摘要：Various extremal results were proved concerning the efficiency of least squares estimates relative to Gauss-Markov estimates and concerning the canonical correlations between 2 sets of random variables. These 2 types of results are related, and a new result is derived from which many previous inequalities can be derived.
-
作者：BARLOW, RE; WU, AS
作者单位：University of California System; University of California Berkeley
摘要：In considering life test experiments from a behavioristic Bayes viewpoint, one may speculate on the effect of model misspecification on the posterior mean under alternative stopping rules. Under the life distribution model, the mean of the posterior is expected to equal the mean of the prior. If the model may be misspecified, this equality is no longer expected. For Bayes estimators of mean life under the exponential model and general priors on mean life, it is expected, based on a preposterio...
-
作者：HALL, P
摘要：Aitchison and Aitken (1976) introduced a novel nonparametric method for estimating probabilities in a multidimensional binary space. The technique is designed for use in multivariate binary discrimination. Their estimator [applied to the data of Anderson et al. (1972) on the diagnosis of keratoconjunctivitis sicca by 10 different symptoms] depends crucially on an unknown smoothing parameter .lambda.; Aitchison and Aitken proposed a maximum likelihood method for determining .lambda. from the sa...
-
作者：PERICCHI, LR
摘要：The analysis of transformation of observations in the linear model with normal errors proposed by Box and Cox (1964) is considered. A different choice of noninformative unnormed prior is advocated, which is not outcome dependent. This new selection of prior leads to a formal identity between likelihood and Bayesian inference, both for the estimation of the best transformation to normality and for the presence of homoscedasticity and additivity under this transformation. Extension to a related ...
-
作者：GEISSER, S
摘要：Predictive sample reuse methods are applied in several ways to the problem of supplying values for the unobserved portion of a vector when N other similar but fully observed vectors are in hand. One method is to use a mixing matrix to to combine 2 independently calculated vector predictors into a single predictor. The other is to use regression techniques with varying constraints on the regression matrix. The methods are applied to 2 data sets [height of boys and weight gain in rats] and for 1...
-
作者：SHIRAHATA, S
作者单位：University of Osaka
摘要：An intraclass rank test of the Spearman type is proposed to test for independence in a bivariate population. The test is useful when there is prior information that the population has the same marginal distributions. The proposed test is asymptotically equivalent to Spearman''s test under the null hypothesis of independence and its contiguous alternatives and is empirically better than that in the case of small sample sizes.
-
作者：TSIATIS, AA
-
作者：BERMAN, M
作者单位：Commonwealth Scientific & Industrial Research Organisation (CSIRO)
摘要：Two classes of point processes, the inhomogeneous and modulated gamma processes, are defined. Asymptotic test statistics of stationarity are derived for modulated gamma processes. Some examples are studied and applied to 2 sets of data [intervals between coal mining disaster and times of arrival of patients to an intensive care unit of a hospital].
-
作者：BESAG, J
摘要：Some exploratory aspects of simple resistant techniques are described and demonstrated on 3 published sets of data. The examples involve resistant regression, resistant analysis of a 2-way table and resistant analysis of a Latin square experiment, respectively. Emphasis is placed on the application of resistant techniques used in conjunction with, and not as a substitute for, classical statistical analysis. [Data on the repellent effect of lime sulfur on the honeybee are discussed].
-
作者：BROWN, BM
摘要：The average of 2 complementary order statistics is termed a symmetric quantile average. In addition to having good efficiency, robustness and computational properties, exact nonparametric confidence intervals for location can be derived from it. A related type of estimator called a combined quantile average is introduced; it is slightly more difficult to compute than the symmetric quantile average, but can have better efficiency and robustness properties. The recommended form of combined quant...
