B1689
Title: Process-based inference for accelerometer and streaming data from wearable devices
Authors: Sudipto Banerjee - UCLA (United States) [presenting]
Pierfrancesco Alaimo Di Loro - LUMSA University (Italy)
Marco Mingione - University of Roma Tre (Italy)
Michael Jerrett - UCLA (United States)
Lipsitt Jonah - UCLA (United States)
Zhou Daniel - UCLA (United States)
Abstract: Rapid developments in streaming data technologies have enabled real-time monitoring of human activity. Wearable devices, such as wrist-worn sensors that monitor gross motor activity (actigraphy or accelerometry), have become prevalent. An actigraph unit (or accelerometer) continually records the activity level of an individual, producing large amounts of high-resolution measurements that can be immediately downloaded and analyzed. While this type of BIG DATA includes both spatial and temporal information, it is argued that the underlying process is more appropriately modelled as a stochastic evolution through time, while accounting for spatial information separately. A key challenge is the construction of valid stochastic processes over paths. A spatial-temporal modelling framework is devised for massive amounts of actigraphy data while delivering fully model-based inference and uncertainty quantification. Building upon recent developments, traditional Bayesian inference is discussed using Markov chain Monte Carlo algorithms as well as faster alternatives such as Bayesian predictive stacking. The methods are tested and validated on simulated data and subsequently, their predictive ability is evaluated on an original dataset from the physical activity through sustainable transport approaches (PASTA-LA) study conducted by UCLA's Fielding School of Public Health.
