The Bayesian estimation of the change-point in independent gaussian samples is considered. The impact of an outlier on the performance of the Bayesian procedure is studied and the posterior density under contamination is given. In this paper, we show that, in the case of small samples, an outlier has a significant impact on the estimation of the change-point. However, if the sample size is enough high, the estimation performs well.

Let $(Y,C,X)$ be a vector of random variables where $Y,$ $C$ and $X$ are, respectively, the interest variable, a right censoring and a covariable (predictor). In this paper, we introduce a new nonlinear wavelet-based estimator of the regression function in the right censorship model. An asymptotic expression for the mean integrated squared error of the estimator is obtained to both continuous and discontinuous curves. It is assumed that the lifetime observations form a stationary $\alpha-$mixing sequence.

Poverty is a vague concept with ill-defined boundaries. And determining a poverty line to divide the population into two classes, poor and non-poor seems steep and hardly capture the complexity of the reality. To overcome this difficulty, we can use fuzzy logic introduced by Zadeh and developed by Dubois and Prade. At the heart of this logic lies the challenge to generalize the classical set theory, assuming there are no precise criteria to distinguish which elements belong to a set or not. This logic is used here to measure poverty in Congo in a multidimensional approach. (paper in French)

We propose a new test of independence of random vectors. We first show that the null hypothesis implies the nullity of the trace of an operator involving inverse regressions covariance operators. Then, using an approach based on slicing, we define a test statistic for which an asymptotic distribution under null hypothesis is derived. Simulations that permit to evaluate the performance of the proposed test with comparisons with existing methods are given.

Several risk measures, such as the distorted insurance premium and the two-sided deviation (TSD) measure, can be regarded as L-functionals with specific weight functions. In this paper, we focus on the TSD risk measure as we define a new estimator by using the bias-reduced estimators of extreme quantiles proposed by Li et al. 2010. A simulation study is carried out to compare, in terms of bias and mean squared error, the new estimator with that introduced recently by Necir and Meraghni 2010.

The contribution of this paper is to produce national smoothed growth indicators that describe the behaviour of economic activity at a monthly frequency, while utilizing a wide range of economic time series. New Eurocoin indicator, that is published monthly by the Bank of Italy, provides a summary index of the medium to long-run component (MLRG) of the gross domestic product (GDP) growth rate and only for the entire aggregate Euro area, by using the Generalized Dynamic Factor Model. The innovation of this research are some procedures, based on Eurocoin approach, to estimate MLRG in countries belonging to Euro Area.

Recently Necir and Meraghni (2009) proposed an asymptotically normal estimator for distortion risk premiums when losses follow heavy-tailed distributions. In this paper, we propose a bias-corrected estimator of this class of risk premiums and establish its asymptotic normality. Our considerations are based on the high quantile estimator given by Matthys and Beirlant 2003.