wavelet-based estimation for bivariate density function using nsdrandom variables

In this paper, we study the asymptotic behavior of the wavelet bivariatedensity function estimator for a negatively superadditive dependent. the convergencerates for the non-linear wavelet estimator are investigated. we evaluate these theoreticalperformances via the minimax approach under the lp risk with p ≥ 1 over a widerange of function classes: the besov classes. under mild assumptions on the model, weshow that it enjoys powerful mean integrated squared error properties.