Efficient Dynamic Prediction of High-density Multilevel Generalized Functional Data

Music City Center 

Dynamic prediction, which typically refers to the prediction of future outcomes using historical records, is often of interest in biomedical research. For datasets with large sample sizes, high measurement density, and multilevel structures, traditional methods are often infeasible because of the computational burden associated with both data scale and model complexity. Moreover, many models do not directly facilitate out-of-sample predictions for multilevel generalized outcomes. To address these issues, we develop a novel approach for dynamic predictions based on a recently developed method estimating complex patterns of variation for exponential family data: Generalized Multilevel Functional Principal Components Analysis (gmFPCA). Our method is able to handle large-scale, high-density multilevel repeated measures much more efficiently, with its implementation feasible even on personal computational resources. The proposed method makes highly flexible and accurate predictions of future trajectories for data that exhibits high degrees of nonlinearity, and allows for out-of-sample predictions to be obtained without reestimating any parameters.

Dynamic Prediction

Functional Data

Longitudinal Data

Wearable Device

Generalized Functional Data

Mixed effect models