1: \begin{abstract}

2: We propose a generalized partially linear functional single index

3: risk score model for repeatedly measured outcomes where the index

4: itself is a function of time. We fuse the nonparametric kernel method

5: and regression spline method, and modify the generalized estimating

6: equation to facilitate estimation and inference.

7: We use local smoothing kernel to estimate the

8: unspecified coefficient functions of time, and use

9: B-splines to estimate the unspecified function of the single index

10: component. The covariance structure is taken into account via a working

11: model, which provides valid estimation and inference procedure whether

12: or not it captures the true covariance. The estimation method is

13: applicable to both continuous and discrete outcomes.

14: We derive large sample properties of the estimation procedure and

15: show a different convergence rate for each component of the model. The

16: asymptotic properties when the kernel and regression spline methods are

17: combined in a nested fashion has not been studied prior to this work,

18: even in the independent data case.

19: \end{abstract}