محمود افشاری

The methods of estimation of nonparametric regression function are quite in statistical application. Using wavelets is one of ways of estimating regression. In this paper, the new mixture prior distri- butions and new bayesian wavelet thresholding estimator of nonparametric regression function are considered. We used the reversible jump algorithm to obtain the appropriate prior distributions and value of thresholding . We surveyed theoretical outcomes with numerical computation and simulation by using R software based on real data. At the end, we compare convergence ratio of given estimator with another by evaluate of average mean square error.