|
Activity Number:
|
158
- SPEED: Statistical Methods, Computing, and Applications Part 2
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 8, 2022 : 10:30 AM to 11:15 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #323748
|
|
|
Title:
|
Analysis of Accelerometer Data from NHANES Database Using Fréchet Single Index Model
|
|
Author(s):
|
Aritra Ghosal * and Wendy Meiring and Alexander Petersen and Marcos Matabuena
|
|
Companies:
|
University of California, Santa Barbara and University of California Santa Barbara and Brigham Young University and University of Santiago de Compostela
|
|
Keywords:
|
Fréchet;
NHANES;
Accelerometer;
Wasserstein ;
Regression ;
Functional
|
|
Abstract:
|
Physical activity improves health and alleviates the causes of mortality. With the advent of accelerometers, we are able to obtain continuous streams of data on metrics of physical activity like step count, amounts of sleep, heart rate, calories spent etc. which helps to understand their distributional nature to predict mortality of the person under study. The distribution of physical activity levels were modeled as objects in L^2 - Wasserstein space and their functional representation could be modeled by Global/Local Fréchet regression or function-on-scalar regression. These models proved to be improvement over summary measures like Total Activity Count (TAC) or other compositional measures as such measures discard important information. Furthermore, the data collected for such studies may follow complicated survey designs which dictates the estimation strategies. This paper aims to build a personalized profile of a patient with the functional representation of activity distribution and model it utilizing the Fréchet Single Index (FSI) model by estimating an index parameter to know its representation advantages. The activity data was obtained from NHANES library.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2022 program
|
