2013/11/29_Leyla Azarang and Irene Castro Conde (University of Vigo)

November 29th 2013

Faculty of Economic and Business Sciences | Salón de Grados

The Jackknife Estimate of Covariance of Two Kaplan-Meier Integrals when Covariables are Present


2013/11/29 – 12:00 h | Leyla Azarang, University of Vigo

Abstract

In this paper the Jackknife estimate of covariance of two Kaplan-Meier integrals with covariates is introduced. Its strong consistency is established under certain conditions.

Adjusted p-values for SGoF multitesting method


2013/11/29 – 12:30 h | Irene Castro Conde, University of Vigo

Abstract

In the field of multiple comparison procedures, adjusted p-values are an important tool to evaluate the significance of a test statistic while taking the multiplicitly issue into account. In this paper we introduce adjusted p-values for Carvajal-Rodríguez et al. (2009)’s Sequential Goodness-of-Fit (SGoF) multitesting method. Main properties of the adjusted p-values are established. Several real data applications are performed to illustrate the practical usage of the adjusted p-values.

SGoF multitesting method under the Bayesian paradigm


2013/11/29 – 13:00 h | Irene Castro Conde, University of Vigo

Abstract

In this paper we consider the Sequential Goodness-of-Fit (SGoF) multitesting procedure under the Bayesian paradigm. For this, it is assumed that the proportion of p-values falling below the significance threshold follows some prior density. Credible intervals and Bayesian preliminary tests for point null hypotheses are combined to define a suitable modification of SGoF method. The performance of Bayesian SGoF is explored through simulations. One of the main conclusions of our research is that the Bayesian viewpoint is suitable to keep the large statistical power of SGoF method even in the presence of strong correlation structures. Application of the method to treatment comparison for acute myocardial infarction and to a microarray study of hereditary breast cancer are included.