On linearization of nonparametric deconvolution estimators for repeated measurements model

By utilizing intermediate Gaussian approximations, this paper establishes asymptotic linear representations of nonparametric deconvolution estimators for the classical measurement error model with repeated measurements. Our result is applied to derive confidence bands density and distribution functions error-free variable interest establish faster convergence rates than ones obtained in existing literature. Due slower decay linearization errors, however, our bootstrap counterparts need be constructed by subsamples.