Vera Afreixo
Adjusted Wald confidence intervals in the meta-analysis of one proportion
Meta-analysis could be used in many fields of research, having high importance in clinical context. In particular, it could be used to estimate the prevalence and incidence of a disease or a multi-resistant bacteria in epidemiological context.
In this work, we discuss the meta-analysis methods for the proportion effect size (e.g. prevalence, incidence), pointing our interest in rare events (analogous for the abundant events) that take into account the practical problem of estimating one low or very low prevalence/incidence.
Several approximate adjusted Wald confidence intervals (CIs) for the meta-analysis of prevalence proportions are proposed. These CIs are developed using a parametric family of shrinkage estimators for estimating the proportions. Two popular statistical models, the fixed-effect model and the random-effects model, are used and discussed in this work. A simulation study is carried out to compare these CIs amongst themselves and compare them with those obtained by the best-known transformation: logit and double arcsine.
This is a joint work with Sara Escudeiro and Adelaide Freitas.
In this work, we discuss the meta-analysis methods for the proportion effect size (e.g. prevalence, incidence), pointing our interest in rare events (analogous for the abundant events) that take into account the practical problem of estimating one low or very low prevalence/incidence.
Several approximate adjusted Wald confidence intervals (CIs) for the meta-analysis of prevalence proportions are proposed. These CIs are developed using a parametric family of shrinkage estimators for estimating the proportions. Two popular statistical models, the fixed-effect model and the random-effects model, are used and discussed in this work. A simulation study is carried out to compare these CIs amongst themselves and compare them with those obtained by the best-known transformation: logit and double arcsine.
This is a joint work with Sara Escudeiro and Adelaide Freitas.
Cláudia Santos
Periodic INAR(1) models based on the signed thinning operator
INteger-valued AutoRegressive (INAR) processes play a central role in the statistical analysis of integer-valued time series. In this work, two INAR (univariate and bivariate) models for the analysis of time series of counts with periodic structure are introduced. Both models are based on the signed thinning operator allowing for positive and negative counts. Basic probabilistic and statistical properties of the periodic models are presented. Innovations are modeled by univariate and bivariate Skellam distributions, respectively. The performance of the conditional least squares and conditional maximum likelihood estimators is compared through a simulation study.
This is a joint work with I. Pereira and M. Scotto.
This is a joint work with I. Pereira and M. Scotto.