Kernel-based random effect time-varying coefficient model for longitudinal data

Lots of efforts have been devoted to develop effective estimation methods for parametric and nonparametric longitudinal data models. Varying coefficient regression model has received a great deal of attention as an important tool for modeling the relation between a response and a group of predictor variables. The varying coefficient model is particularly useful in longitudinal data analysis. A random effect time-varying coefficient model is proposed for analyzing longitudinal data, which is based on the basic principle of least squares support vector machine along with the kernel technique. A generalized cross validation method is also considered for choosing the tolerance level and the hyperparameters which affect the performance of the proposed model. The proposed model is evaluated through numerical studies.