Abstract:
|
Multivariate varying coefficient model with functional response has become an important statistical tool for many neuroimaging studies. In this paper, we study estimation of varying coefficient functions and variable selection simultaneously. In the ultra-high dimensional setting, we investigate the minimax optimal rate and model selection property at the same time under both fixed and random designs. The algorithm based on the ADMM is developed to obtain the estimator. The finite-sample performance is demonstrated through both simulation and real data analysis.
|