Activity Number:
|
481
- Nonparametric Methods in Functional Data Analysis
|
Type:
|
Contributed
|
Date/Time:
|
Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
|
Sponsor:
|
Section on Nonparametric Statistics
|
Abstract #309720
|
|
Title:
|
Common Latent Factor Analysis for Functional and Multivariate Data
|
Author(s):
|
Joonho Gong* and Luo Xiao and Arnab Maity
|
Companies:
|
North Carolina State University and North Carolina State University and North Carolina State University
|
Keywords:
|
Functional data analysis;
Penalized splines;
Dimension reduction;
Principal components
|
Abstract:
|
We propose a new Principal Component Analysis to jointly analyze functional and multivariate data. This approach identifies common latent factors shared in the two types of data by using the cross-covariance function, which allows to decompose the data into shared components and independent ones. The decomposition not only reveals the latent structure with respect to the variability that can be explained by each component but also implements dimension reduction by mapping the original data into a low-dimensional representation. Through a simulation study, we demonstrate this method is robust to high measurement error and relatively large variance of the independent components. Application to real data is illustrated to show how the proposed approach contributes to simultaneous analysis of both data.
|
Authors who are presenting talks have a * after their name.