|
Abstract:
|
Wearable devices provide a compelling framework for understanding circadian rhythms and diurnal patterns of activity. Using methods for separating amplitude and phase variability in exponential family functional data, we develop an approach to uncover the distinct phenotypes, or chronotypes, that give rise to differences in these patterns in physical activity. Our method for aligning or registering observed activity data alternates between two steps: the first uses generalized functional principal components analysis (GFPCA) to calculate template functions, and the second estimates smooth warping functions that map observed curves to templates. The results of both steps provide unique, data-driven insights into the processes behind activity. Our motivation comes from the Baltimore Longitudinal Study on Aging, in which accelerometer data provides observations of activity and sedentary behavior. We analyze binary functional data with observations each minute over 24 hours for 592 participants, where values represent activity and inactivity.
|
