Block thresholding in wavelet estimation of the regression function with errors in variables

Errors-in-variables regression is the study of the association between covariates and responses where covariates are observed with errors. In this paper, we consider the estimation of regression functions when the independent variable is measured with error. We investigate the performances of an adaptive wavelet block thresholding estimator via the minimax approach under the Lp risk with p ≥ 1 over Besov balls. We prove that it achieves the optimal rates of convergence.