| ¡@¡@We propose a nonlinear wavelet shrinkage estimator. Such an estimator combines the asymptotic equivalence to the best linear unbiased prediction (BLUP) and the Bayesian estimation. Some optimality properties are discussed. Characterization of prior and posterior regularity is also discussed. The smoothness of proposed estimators can be controlled via tuning parameters. A data-driven GCV method is used to select these parameters. A simulation study is carried out for comparison with other methods. (This talk is based on joint works with Henry Horng-Shing Lu, Institute of Statistics, National Chiao Tung University.) |