Spline estimation of functional principal components via manifold conjugate gradient algorithm

DOI10.1007/S11222-022-10175-2zbMATH Open1499.62020arXiv2211.04784OpenAlexW4308749841MaRDI QIDQ2103990FDOQ2103990

Authors: Shiyuan He, Hanxuan Ye, Kejun He 

Publication date: 9 December 2022

Published in: Statistics and Computing (Search for Journal in Brave)

Abstract: Functional principal component analysis has become the most important dimension reduction technique in functional data analysis. Based on B-spline approximation, functional principal components (FPCs) can be efficiently estimated by the expectation-maximization (EM) and the geometric restricted maximum likelihood (REML) algorithms under the strong assumption of Gaussianity on the principal component scores and observational errors. When computing the solution, the EM algorithm does not exploit the underlying geometric manifold structure, while the performance of REML is known to be unstable. In this article, we propose a conjugate gradient algorithm over the product manifold to estimate FPCs. This algorithm exploits the manifold geometry structure of the overall parameter space, thus improving its search efficiency and estimation accuracy. In addition, a distribution-free interpretation of the loss function is provided from the viewpoint of matrix Bregman divergence, which explains why the proposed method works well under general distribution settings. We also show that a roughness penalization can be easily incorporated into our algorithm with a potentially better fit. The appealing numerical performance of the proposed method is demonstrated by simulation studies and the analysis of a Type Ia supernova light curve dataset.

Full work available at URL: https://arxiv.org/abs/2211.04784

Recommendations

• Functional principal components analysis via penalized rank one approximation
• Parametric Functional Principal Component Analysis
• Principal components analysis of sampled functions
• Bayesian estimation of principal components for functional data
• Penalized PCA approaches for B-spline expansions of smooth functional data

zbMATH Keywords

functional data analysisStiefel manifoldpositive definite coneproduct manifoldmatrix Bregman divergence

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Factor analysis and principal components; correspondence analysis (62H25) Functional data analysis (62R10) Statistics on manifolds (62R30)

Cites Work

• Functional data analysis.
• Principal component models for sparse functional data
• Functional Clustering and Identifying Substructures of Longitudinal Data
• Functional Data Analysis for Sparse Longitudinal Data
• Title not available (Why is that?)
• Applied functional data analysis. Methods and case studies
• The Geometry of Algorithms with Orthogonality Constraints
• Title not available (Why is that?)
• Title not available (Why is that?)
• Function minimization by conjugate gradients
• Title not available (Why is that?)
• Functional linear regression analysis for longitudinal data
• Properties of principal component methods for functional and longitudinal data analysis
• From sparse to dense functional data and beyond
• Uniform convergence rates for nonparametric regression and principal component analysis in functional/longitudinal data
• Smoothed functional principal components analysis by choice of norm
• Clustering for Sparsely Sampled Functional Data
• A statistical perspective on ill-posed inverse problems (with discussion)
• Title not available (Why is that?)
• On the joint convexity of the Bregman divergence of matrices
• Functional linear discriminant analysis for irregularly sampled curves
• Optimization methods on Riemannian manifolds and their application to shape space
• Consistency of restricted maximum likelihood estimators of principal components
• Conic Geometric Optimization on the Manifold of Positive Definite Matrices
• Fast Computation of Fully Automated Log-Density and Log-Hazard Estimators
• Matrix Nearness Problems with Bregman Divergences
• Covariance, subspace, and intrinsic Crame/spl acute/r-Rao bounds
• Title not available (Why is that?)
• Title not available (Why is that?)
• A brief introduction to manifold optimization
• Asymptotic properties of penalized spline estimators in concave extended linear models: rates of convergence

Cited In (1)

• Penalized spline estimation of principal components for sparse functional data: rates of convergence

Uses Software

• fda (R)
• fpca

This page was built for publication: Spline estimation of functional principal components via manifold conjugate gradient algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2103990)