Abstract:This study proposes a method to incorporate covariate information into sparse functional data analysis. The method aims at cases where each subject has a limited number of longitudinal measurements and is associated with static covariates. This research is motivated by several use cases in practice. One representative example is void swelling, a nuclear-specific material degradation mechanism. Void swelling is affected by many covariates, including alloy composition and irradiation type. How to accurately model the complicated joint effects of such covariates on the swelling process is the key to mitigating the effect of swelling and ensuring safe operation. Unlike most of the existing methods, the proposed method can handle high-dimensional covariates with the informative covariate identification procedure and sparse and irregularly spaced measurements, that is, does not require complete or dense observations. The main innovation of the proposed method is that we model the variation coming from covariates and the variation left conditioned on covariates, such that the functional principal component analysis and Gaussian process can be conducted in a unified manner. We also propose a systematic approach to identify important covariates in the hypothesis testing context. The methodology is demonstrated on applications in nuclear engineering and healthcare and simulation studies. Funding: This work was supported in part by the Air Force Office of Scientific Research [Grant FA9550-20-1-0072] and in part by the Department of Energy [Award DE-NE0008993]. Data Ethics & Reproducibility Note: No data ethics considerations are foreseen related to this article. The code capsule is available on Code Ocean at https://codeocean.com/capsule/9363709 and in the e-Companion (available at https://doi.org/10.1287/ijds.2023.0025 ).

Adaptive Functional Thresholding for Sparse Covariance Function Estimation in High Dimensions

Adaptive Thresholding for Sparse Covariance Matrix Estimation

Covariance Operator Estimation via Adaptive Thresholding

Fast covariance estimation for multivariate sparse functional data

Fast Covariance Estimation for Sparse Functional Data

Covariance Function Estimation for High-Dimensional Functional Time Series with Dual Factor Structures

Mean and Covariance Estimation for Discretely Observed High-Dimensional Functional Data: Rates of Convergence and Division of Observational Regimes

Factor-guided estimation of large covariance matrix function with conditional functional sparsity

Large Covariance Estimation by Thresholding Principal Orthogonal Complements

Sparse Functional Principal Component Analysis in High Dimensions

Framelet Block Thresholding Estimator for Sparse Functional Data

Nonparametric Operator-Regularized Covariance Function Estimation for Functional Data

High dimensional test for functional covariates

Doubly Functional Graphical Models in High Dimensions

Adaptive estimation of irregular mean and covariance functions

Minimax Optimal Estimation of High-Dimensional Sparse Covariance Matrices with Missing Data.

Large-scale Multiple Testing of Cross-covariance Functions with Applications to Functional Network Models

Covariate Dependent Sparse Functional Data Analysis

On estimation in varying coefficient models for sparse and irregularly sampled functional data

CANONICAL THRESHOLDING FOR NON-SPARSE HIGH-DIMENSIONAL LINEAR REGRESSION

Oracle-Efficient Confidence Envelopes for Covariance Functions in Dense Functional Data