Alba, M.V. Barrera, D. Jiménez, M.D.
Published in
Computational Statistics

In this paper, a test for the homogeneity of two populations is proposed. It is based on the L2-norm of the difference of the empirical characteristic functions associated to two independent random samples of each population. A quadrature formula is used to construct the test function by using the cubic many-knot Hermite spline interpolation. In or...

Costa Mattos, Néli Maria dos Santos Migon, Hélios
Published in
Computational Statistics

In this paper MCMC (Markov Chain Monte Carlo) techniques are proposed to perform Bayesian inference to evaluate the reliability of units with Weibull lifetime, submitted to accelerated and censored life tests. A full Bayesian analysis is done via Gibbs sampling. The marginal posterior of the main parameters and other unobserved quantities of intere...

Louzada-Neto, Francisco
Published in
Computational Statistics

We develop an approximate Bayesian analysis for hazard models with shape parameters dependent on covariates. We consider a general hazard regression model which includes, among others, the proportional hazards and the accelerated failure time models, with the inverse power law and the Arrhenius models as relationship of the scale parameter and a co...

Kim, Jinhyo Seo, Sangwon
Published in
Computational Statistics

A numerical quadrature of a particular probability integral is concerned with using the Fourier transformation which smoothes the stiffness. The Fourier transformation of the Laplace distribution becomes, in a statistical sense, the Cauchy distribution. It is shown that the Gauss-Hermite quadrature of the Cauchy distribution, equivalent to the Four...

Heng-Hui, Lue
Published in
Computational Statistics

A new method for nonparametric regression data analysis by analyzing the sensitivity of normally large perturbations with the Principal Hessian Directions (PHD) method (Li 1992) is introduced, combining the merits of effective dimension reduction and visualization. We develop techniques for detecting perturbed points without knowledge of the functi...

Marron, J.S. Chung, S. S.
Published in
Computational Statistics

The product of most statistical smoothing methods is a single curve estimate. A drawback of such methods is that what is learned varies with choice of the smoothing parameter. This paper proposes simultaneous display of all important features, through presentation of a family of smooths. Some suggestions are given as to how this should be done.

van der Linde, Angelika
Published in
Computational Statistics

Estimation of a smooth predictor function in logistic regression requires the determination of a smoothing parameter. Several cross-validatory criteria for finding such a smoothing parameter have been proposed generalizing techniques that are asymptotically well performing for Gaussian data. Here it is argued that a smoothing parameter is a model p...

Linton, Oliver
Published in
Computational Statistics

Chamayou, J.-F.
Published in
Computational Statistics

The Chambers, Mallows and Stuck algorithm for stable pseudo random numbers is applied to the generation of Landau variates The infinitely divisibility property of the Vavilov density is used to generate the variates. Use is made of the convolution between a Vavilov density with velocity β and the density of the sum of an increasing number of produc...

Cheng, Jian
Published in
Computational Statistics

We show two distribution-independent algorithms to estimate the mean of bounded random variables, one with the knowledge of variance, the other without. These algorithms guarantee that the estimate in within the desired precision with an error probability less than or equal to the requirement. Some simplified stopping rules are also given.