Asymptotic Properties Of Semiparametric Models Based On Negatively Associated Samples

The semiparametric regression models is one of the most widely used models inmodern statistics. In many semiparametric regression models, we always consider theerror is independent and identically distributed. However, in many practical problems,errors often show some associated. So it is very important theoretical and practicalsignificance to study semiparametric regression model with negative associatederrors.In chapter2, using the general weight function and least squares estimationmethod, we get the least squares estimators(LSE) and the weighted least squaresestimators(WLSE) of parameter, nonparametric parts and error variance. We get theirweak consistency under some suitable conditions.In chapter3, we get the least squares estimators of parameter and nonparametricparts. Under the suitable conditions, we obtain the asymptotic normality results fortheir LSE.