Making sense of wearable accelerometer data by Vincent van Hees

This eScience talk was presented at

4th National eScience Symposium
Sports & eHealth Session
13 October 2016, Amsterdam ArenA
https://www.esciencecenter.nl/event/program/4th-national-escience-symposium/

SPEAKER: Vincent van Hees – Netherlands eScience Center

TALK: Making sense of wearable accelerometer data collected under uncontrolled real life conditions

ABSTRACT: Measuring human physical activity and sleep is crucial in health research and increasingly performed with wrist-worn acceleration sensors. However, robust classification of detailed activity types and estimation of energy expenditure remains difficult under uncontrolled real life conditions. Our aim is to extract insight from the sensor data beyond the results from current heuristic algorithms, while trying to avoid over-interpretation given the uncertainties that are introduced by uncontrolled real life experimental conditions. We developed an unsupervised data-driven model for identifying clusters in the accelerometer time series data. To aid interpretation we fuse the model output with activity diary recordings as well as conventional heuristic algorithm output.

ABOUT: Vincent holds a PhD in Epidemiology from the University of Cambridge and did a post-doc at Newcastle University. Central theme of Vincent’s work has been the development of algorithms to process data from wearable movement sensors as used for population research on human behaviour. At the Netherlands eScience Center, Vincent’s current focus is on novel approaches for time series and sensor data analysis. Vincent published several journal articles on algorithms for automatic interpretation of movement sensor data. His methods are available as generic open source software (R package GGIR), which is increasingly used by the research community.

source

source https://gadgetsteam.com/2019/05/12/making-sense-of-wearable-accelerometer-data-by-vincent-van-hees/