Large Sample Properties Of The Regression Modle Of Half-Paramter Under Error Being Martingale Difference

In recent years, whether in theoretical research or practical application, the semiparametric regression model has been highly regarded of many scholars. Be-cause of the semiparametric regression model is a mixture between parameters and the parameters of two kinds of single model, so it has more flexibility that compared with the linear model or nonparametric model. In practical problem-S,Semiparametric regression model is a modle of what more close to reality what more make full use of the information contained in the data. And then,it is certain theoretical value and practical significance to use the model in real life instead of a single nature regression model to consider the practical problems.Under the basis of existing research, this paper discussed some of the large sample properties of the estimator of semiparametric regression model with mar-tingale difference erros. Based on nonparametric estimator, construct least squares estimator and weighted least squares estimator of the parameters, and under the basis of conditions, deduced the their moment consistency and asymptotic nor-mality.