目前纵向数据动态模型统计推断的研究刚刚起步，存在许多亟待解决的问题。本课题主要研究纵向数据动态非参数模型的统计推断，拟研究的主要内容有：.（1）基于局部线性光滑技术和self-normalization方法，研究纵向数据动态非参数模型在稀疏/密集情形下的一致估计问题。（2）基于局部线性估计、受约束的紧积 (tensor product) B 样条和广义估计方程方法，研究纵向数据动态非参数模型部分个体观测次数趋于无穷时非参数函数的有效估计问题。.希望本课题的研究能对纵向数据动态非参数回归模型的统计推断及其应用领域的发展做出一定的贡献，同时为下一步动态半参数模型密集/稀疏纵向数据的一致估计研究和部分个体观测次数趋于无穷时动态半参数模型的研究奠定坚实的理论基础。

The worldwide authors have paied much attention on the statistical inference of longitudinal data time-dynamic model. The proposal aims to investigate the following statistical inference issues of longitudinal data time-dynamic nonparametric models: (1) Based on the local linear smoothing technique and self-normalization methods, we will investigate a unified inference for sparse and dense longitudinal time-dynamic nonpara metric model; (2) Based on the local linear regression, tensor product B-spline and generalized estimating equation, we will construct efficient estimation for longitudinal time-dynamic nonparametric model when the numbers of observation for some subjects tend to infinity. We hope that our researches could make some contributions to the statistical inference and related application area of longitudinal time-dynamic nonparametric models. And also lay a strong foundation for the further development of sparse and dense longitudinal time-dynamic semi-parametric model's unified inference.