ديهمت جذومن رادحنلاا يملعملالا مادختسأب بولسأ زيب

Abstract

This research is aimed to smooth the nonparametric model by applying Bayesian approach in estimation where this approach has accurate results through minimizing estimation of loss function. This approach is achieved by considering a binary variable whose value controls the process of choosing of variables. The Gray code is dependent in updating the stream of binary variable while the focused sampling is dependent in algorithm of smoothing of 2nonparametric regression function. We deduced: (1) Gray code is efficient in this application but it needs complete scan to space of variables, (2) the focused sampling takes essential role in focusing on the active variables in each iterationand in partitioning the variables in subsets to simplify procedure of smoothing. We get very good results of plotting of regression also good values for criterions of errors.