Showing papers in "Computational Statistics & Data Analysis in 1996"

TL;DR: This paper compares partially parametric and fully parametric regression-based multiple-imputation methods for handling data sets with missing values and provides an example of how multiple imputation can be used to combine information from two cohorts to estimate quantities that cannot be estimated directly from either one of the cohorts separately.

TL;DR: An exact algorithm for computing depth contours of a bivariate data set using the half-space depth introduced by Tukey is constructed and it is shown how depth contour can be used to construct robustified versions of classification techniques based on convex hulls.

TL;DR: In this article, a false discovery rate (FDR) based thresholding procedure was proposed to control the expected proportion of incorrectly included coefficients among those chosen for the wavelet reconstruction, which is inherently adaptive and responds to the complexity of the estimated function and to the noise level.

TL;DR: This paper presents an EM algorithm for computing maximum-likelihood estimates of the parameters of a Markov-modulated Poisson process, and compares it to the Nelder-Mead downhill simplex algorithm.

TL;DR: In this article, seven tests of equality of variances are compared in terms of robustness and power in a simulation experiment with small-to-moderate sample sizes, where data are assumed to come from a location-scale family with unknown means, variances, and density functions.

TL;DR: A comparative investigation of both logistic regression models and feed-forward neural networks including some extensions is presented and the theoretical features and properties are reviewed and illustrated in two examples.

TL;DR: It is shown how suitable clustering criteria or grouping methods may be derived from probabilistic models for partition-type, hierarchical and tree-like clustering structures in the case of vector-valued data, dissimilarity matrices and similarity relations.

TL;DR: The aim of this paper is to compare three methods based on the hypervolume criterion with other well-known methods for determining the number of clusters by pointing out the performance of each and giving some recommendations to help potential users of these techniques.

TL;DR: In this article, a spatio-temporal stochastic process is used to predict the snow water equivalent (SWE) of the Animas River basin in southwest Colorado, where the US National Weather Service used a purely spatial model to predict SWE at sites where no observations are available.

TL;DR: In this paper, three alternative estimation procedures based on the EM algorithm are considered, two of them make use of numerical integration techniques (Gauss-Hermite or Monte Carlo), and the third one is a EM type algorithm based on posterior modes.

TL;DR: This article proposed a method for simultaneous variable selection and outlier identification based on the computation of posterior model probabilities, which avoids the problem that the model selection depends upon the order in which variable selection is carried out.

TL;DR: It is argued that many statistical and mathematical restrictions that usually restrict modeling and analysis can be dispensed with by employing the GA as an optimization technique.

TL;DR: The paper provides a survey of work in constrained classification, in which constraints restrict the set of allowable solutions, and ways of assessing the results of a constrained classification study are surveyed.

TL;DR: In this paper, a zero adjusted discrete model is developed, where the proportion of zeros in the data is higher (lower) than that predicted by the original model, and the effect of such an adjustment is studied.

TL;DR: In this article, a Monte-Carlo analysis of the effect of the cutoff point on the difference between the two groups is presented, showing that the optimal cutoff value results in an overestimation of the differences between the prognostic groups.

TL;DR: This paper presents a collection of macros for The SAS System to perform meta-analyses of clinical trials where the results of a single trial can be displayed in a fourfold-table.

TL;DR: In this paper, a data dependent technique for selecting a threshold with which to shrink empirical wavelet coefficients is introduced, based on standard statistical tests of hypotheses, which gives good results both when the underlying function is constant, and when it undergoes multiple abrupt changes.

TL;DR: In this paper, the authors used spectral analysis for the exploratory analysis of spatial point patterns, relating periodogram structure to the type of stochastic process which could have generated an observed pattern.

TL;DR: Several methods are considered which incorporate cluster-specific scatter matrices which enables them to describe elliptical clusters with different orientation, and the distinction between these methods lies in the way they deal with clusters of different volume, cardinality, and density.

TL;DR: The coefficient of variation (CV) is a relatively dimensionless measure of variability (relative standard deviation) as mentioned in this paper, a parameter of major importance in clinical and diagnostic areas, but it is rarely mentioned in statistical literature, and the reason for this is that these functions are not offered by standard statistical software.

TL;DR: In this article, the authors discuss some practical aspects of fitting and interpreting nonlinear quasi-likelihood regression models, including models with a parametric link function, as well as models having a link to a nonlinear predictor having multiplicative terms.

TL;DR: In this paper, the problem of estimating the data variance as well as the parameter vector via an extended least-squares technique motivated by maximum likelihood estimation is considered and convergence of an algorithm that generalizes a standard successive approximation algorithm from nonlinear programming is shown.

TL;DR: In this paper, the authors describe the prediction variance distribution on a sphere by using a plot of its quantiles, and a method for deriving these quantiles is given to illustrate the utility of the proposed approach.

TL;DR: In this paper, the authors proposed a one-sided studentized range test (OSRT) for testing the null hypothesis H0:?1 = … =?k against the simple ordered alternative Ha:? 1? …?k in a one way layout.

TL;DR: In this article, an unconditional expected squared error criterion is used for an overall comparison of five different prediction methods: Principal component regression by the size of the eigenvalues (PCR1), partial least squares regression (PLS), restricted principal component regression (RPCR), and modified maximum likelihood regression (MML).

TL;DR: In this article, the problem of identifying multiple abrupt change points in a sequence of observations is approached via hypothesis testing, and the null hypothesis of no change points is tested and if it is rejected the number of change points present are indicated by the procedures introduced.

TL;DR: Nguyen and Miller as discussed by the authors introduced the essential features of TS, applied TS to the problem of constructing an exact D-optimal design for a main effect or a quadratic model with a finite design space, and compared performances of TS and the Fedorov exchange algorithm as modified by Nguyen and Miller.

TL;DR: This paper considers the issue of exploration in theory and in practice in the form of a simple example, which enables us to identify general properties of the exploration types, and to comment about when exploration would be profitable.

TL;DR: In this article, the results of a Monte Carlo study of the size and power of parametric and semi-parametric approaches to inference on covariate effects in survival (time-to-events) models in the presence of model misspecification and an independent censoring mechanism are reported.