|
Abstract:
|
Personal wearable medical devices are becoming increasingly useful in medical care. Many of these devices are marked by automatic high-frequency data collection and are used to track physiological variables, behavioral patterns, and clinical symptoms. This emerging field of tracking data analysis calls for urgent development of theory, methods, and algorithms in order to improve the existing scope of knowledge and applications for better measurement outcomes and industrial products. The long-term aim of this study is to develop novel statistical methods and fast computational algorithms for analyzing high frequency temporal tracking data that are more robust and accurate than current methodologies, and to identify and understand patterns of activity and their changes over space and time. The first aim of this project focuses on analyzing accelerometer data using wavelet decomposition analysis. Wavelet decomposition is a useful statistical tool to analyze functional components of longitudinal data, while maintaining temporal information. By conducting wavelet decomposition on time series data, one can achieve dimension reduction and maintain the important time-specific information.
|
