Two sample nonparametric procedures based on sample coverages for uncensored data.
Two-Sample Rank Tests with Truncated Populations.
Un algorithme de calcul de la probabilité qu'une variable de Fisher-Snedecor, à premier degré de liberté impair, excède une valeur donnée
Un test non paramétrique unilatéral de rupture d'homogénéité de «k» échantillons
Una caracterización de la distribución de Cauchy.
Las f.d.'s (funciones de distribución) de Cauchy tienen un puesto importante en la historia moderna de probabilidad y estadística. También, esas f.d.'s son corrientemente de interés en estudios de "robustez" de varios estadísticos.En esta nota se quiere dar una caracterización sencilla de las distribuciones de Cauchy, y unas ideas sobre la independencia de combinaciones lineales de variables i.i.d. (independientes e idénticamente distribuidas) con una f.d. común de Cauchy.En los trabajos de Lukacs...
Una interpretación geométrica del equilibrio para variables aleatorias usuales
Une caractérisation des lois exponentielles et géométriques
Une famille de distributions : des paretiennes aux «contra-paretiennes». Applications à l'étude de la concentration urbaine et de son évolution
Ce texte est consacré à une famille de distributions statistiques — qui comprend les distributions de V. Pareto, celles du type exponentiel et celles que l'on appellera ici «contra-paretiennes» (ou «anti-paretiennes») — dont l'unité tient à ce qu'elles vérifient toutes une même relation fonctionnelle. Celle-ci est d'ailleurs interprétable en termes d'inégalité des distributions ; elle fournit en outre une méthode simple et efficace d'ajustement des distributions de la famille à des «données» observées....
Une mesure de la déviation quadratique d'estimateurs non paramétriques
Une remarque sur l'article de M. A. Guldberg: “On discontinuons frequency functions and statistical series.”
Uniform rates of convergence in extreme-value theory. Normal and gamma models
Univariate parametric survival analysis using GS-distributions.
The GS-distribution is a family of distributions that provide an accurate representation of any unimodal univariate continuous distribution. In this contribution we explore the utility of this family as a general model in survival analysis. We show that the survival function based on the GS-distribution is able to provide a model for univariate survival data and that appropriate estimates can be obtained. We develop some hypotheses tests that can be used for checking the underlying survival model...
Variance of periodic measure of bounded set with random position
The principal term in the asymptotic expansion of the variance of the periodic measure of a ball in under uniform random shift is proportional to the st power of the grid scaling factor. This result remains valid for a bounded set in with sufficiently smooth isotropic covariogram under a uniform random shift and an isotropic rotation, and the asymptotic term is proportional also to the -dimensional measure of the object boundary. The related coefficients are calculated for various periodic...
Weak convergence of linear rank statistics
Weighting, likelihood ratio order and life distributions
We use weighted distributions with a weight function being a ratio of two densities to obtain some results of interest concerning life and residual life distributions. Our theorems are corollaries from results of Jain et al. (1989) and Bartoszewicz and Skolimowska (2006).
What is the best approximation of ruin probability in infinite time?
We compare 12 different approximations of ruin probability in infinite time studying typical light- and heavy-tailed claim size distributions, namely exponential, mixture of exponentials, gamma, lognormal, Weibull, loggamma, Pareto and Burr. We show that approximation based on the Pollaczek-Khinchin formula gives most accurate results, in fact it can be chosen as a reference method. We also introduce a promising modification to the De Vylder approximation.
Wild bootstrap in RCA(1) model
In the paper, a heteroskedastic autoregressive process of the first order is considered where the autoregressive parameter is random and errors are allowed to be non-identically distributed. Wild bootstrap procedure to approximate the distribution of the least-squares estimator of the mean of the random parameter is proposed as an alternative to the approximation based on asymptotic normality, and consistency of this procedure is established.
Wrapped Distributionsand Measurement Errors.
Zur Schätzung eines Dichtefunktionals.
как критерий независимости в таблице сопряженности признаков