中文摘要：

本文考虑纵向数据半参数回归模型，通过考虑纵向数据的协方差结构，基于Profile最小二乘法和局部线性拟合的方法建立了模型中参数分量、回归函数和误差方差的估计量，来提高估计的有效性，在适当条件下给出了这些估计量的相合性．并通过模拟研究将该方法与最小二乘局部线性拟合估计方法进行了比较，表明了Profile最小二乘局部线性拟合方法在有限样本情况下具有良好的性质．

英文摘要：

In this paper, we study the semiparametric regression model for longitudinal data. Through considering the covariance structure of longitudinal data and using the profile least squares local linear fitting method to improve the efficiency of the estimators, we obtain the estimators of parametric component, nonparametric function and error vari- ance. Under the suitable conditions we give their consistency and compare this method with the least squares local linear fitting method through a Monte Carlo simulation studies which illustrates the mothed has good properties under the finite - sample condition.