-
作者：NEWTON, MA; RAFTERY, AE
作者单位：University of Wisconsin System; University of Wisconsin Madison; University of Washington; University of Washington Seattle
摘要：We introduce the weighted likelihood bootstrap (WLB) as a way to simulate approximately from a posterior distribution. This method is often easy to implement, requiring only an algorithm for calculating the maximum likelihood estimator, such as iteratively reweighted least squares. In the generic weighting scheme, the WLB is first order correct under quite general conditions. Inaccuracies can be removed by using the WLB as a source of samples in the sampling-importance resampling (SIR) algorit...
-
作者：SEVERINI, TA
摘要：Let Y1, . . ., Y(n) denote independent observations of the form Y(j) = theta + sigmaepsilon(j) where epsilone1, . . ., epsilon(n) are independent random variables each distributed according to a density p, and theta and sigma are unknown parameters. This paper presents a method for conditional inference about the location parameter theta that does not require the specification of a parametric family of densities for p. This method uses the configuration statistic to select the density p and is...
-
作者：RIPLEY, BD
摘要：Feed-forward neural networks are now widely used in classification problems, whereas non-linear methods of discrimination developed in the statistical field are much less widely known. A general framework for classification is set up within which methods from statistics, neural networks, pattern recognition and machine learning can be compared. Neural networks emerge as one of a class of flexible non-linear regression methods which can be used to classify via regression. Many interesting issue...
-
作者：LEWIS, SM; DEAN, AM; DRAPER, NR; PRESCOTT, P
作者单位：University of Wisconsin System; University of Wisconsin Madison; University System of Ohio; Ohio State University
摘要：In industrial experimentation designs are needed to estimate the effects of mixture components independently of process variables. This paper gives general methods of constructing designs for q greater-than-or-equal-to 3 mixture components in two or more orthogonal blocks. The designs are built from Latin squares with specific properties and are generalizations of those for four and five components given by Draper and co-workers and Prescott and co-workers respectively.
-
作者：WHITTLE, P; KAY, J; HAND, DJ; TARASSENKO, L; BROWN, PJ; TITTERINGTON, DM; TAYLOR, C; GILKS, WR; CRITCHLEY, F; MAYNE, AJ; WAHBA, G; LUTTRELL, SP; BACZKOWSKI, AJ; MARDIA, KV; BREIMAN, L; BUNTINE, W; CHATFIELD, C; DEVEAUX, RD; DARKEN, CJ; UNGAR, LH; GLENDINNING, RH; HASTIE, T; TIBSHIRANI, R; MCLACHLAN, GJ; MICHIE, D; OWEN, AB; WOLPERT, DH; RIPLEY, BD
作者单位：University of Edinburgh; The Santa Fe Institute; University of Oxford; Princeton University; Siemens AG; Siemens USA; University of Stirling; Open University - UK; University of Toronto; University of Queensland; University of Liverpool; University of Glasgow; University of Leeds; Stanford University; MRC Biostatistics Unit; University of Birmingham; University of Wisconsin System; University of Wisconsin Madison; University of California System; University of California Berkeley; University of Pennsylvania; Nokia Corporation; Nokia Bell Labs; AT&T; University of Bath
-
作者：PEARCE, SC
摘要：Reinforcement is here meant the addition of plots of standard treatments to the blocks of a non-orthogonal design to improve it. The purpose could be either to aid a comparison of the existing treatments or to facilitate their comparison with the standards. The position is examined for both square and rectangular lattices in block designs and for rectangular lattices in row-and-column designs. It is concluded that the method could be useful in practice.
-
作者：GRENANDER, U; MILLER, MI
作者单位：Brown University; Washington University (WUSTL)
摘要：Modern sensor technologies, especially in biomedicine, produce increasingly detailed and informative image ensembles, many extremely complex. It will be argued that-pattern theory can supply mathematical representations of subject-matter knowledge that can be used as a basis for algorithmic 'understanding' of such pictures. After a brief survey of the basic principles of pattern theory we shall illustrate them by an application to a concrete situation: high magnification (greater than 15000 x)...
-
作者：CONSTANTINE, AG; HALL, P
作者单位：Australian National University
摘要：The fractal dimension D of stationary Gaussian surfaces may be expressed very simply in terms of behaviour of the covariance function near the origin. Indeed, only the covariance of line transect samples is required, and that fact makes practical estimation of D relatively straightforward. The case of non-Gaussian surfaces is more poorly understood, but we might define 'effective fractal dimension' in terms of the covariance function, as though the surface were Gaussian. In the present paper w...
-
作者：MAGULURI, G; ZHANG, CH
摘要：The proportional mean residual life model was originally proposed by Oakes and Dasu. This model can be extended to a regression model with explanatory variables. We consider the problem of estimation of the regression parameter in this general model. We investigate two types of estimator, of which one is based on the maximum likelihood equation of the exponential regression model, and the other is based on the underlying proportional hazards structure of the model and Cox's estimating equation...
-
作者：MARRON, JS; RUPPERT, D
作者单位：Cornell University; University of North Carolina; University of North Carolina Chapel Hill
摘要：We consider kernel estimation of a univariate density whose support is a compact interval. If the density is non-zero at either boundary, then the usual kernel estimator can be seriously biased. 'Reflection' at a boundary removes some bias, but unless the first derivative of the density is 0 at the boundary the estimator with reflection can still be much more severely biased at the boundary than in the interior. We propose to transform the data to a density that has its first derivative equal ...
