Wavelet linear estimation of a density and its derivatives from observations of mixtures under quadrant dependence

Abstract. The estimation of a density and its derivatives from a finite mixture under the pairwise positive quadrant dependence assumption is considered. A new wavelet based linear estimator is constructed. We evaluate its asymptotic performance by determining an upper bound of the mean integrated squared error. We prove that it attains a sharp rate of convergence for a wide class of unknown densities.