Fast covariance estimation for high-dimensional functional data

DOI10.1007/S11222-014-9485-XzbMATH Open1342.62094DBLPjournals/sac/XiaoZRC16arXiv1306.5718OpenAlexW2070330801WikidataQ31048350 ScholiaQ31048350MaRDI QIDQ2631375FDOQ2631375

Authors: Luo Xiao, Vadim Zipunnikov, Ciprian M. Crainiceanu, David Ruppert 

Publication date: 29 July 2016

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

Abstract: For smoothing covariance functions, we propose two fast algorithms that scale linearly with the number of observations per function. Most available methods and software cannot smooth covariance matrices of dimension $JimesJ$ with $J>500$; the recently introduced sandwich smoother is an exception, but it is not adapted to smooth covariance matrices of large dimensions such as $Jge10,000$. Covariance matrices of order $J=10,000$, and even $J=100,000$, are becoming increasingly common, e.g., in 2- and 3-dimensional medical imaging and high-density wearable sensor data. We introduce two new algorithms that can handle very large covariance matrices: 1) FACE: a fast implementation of the sandwich smoother and 2) SVDS: a two-step procedure that first applies singular value decomposition to the data matrix and then smoothes the eigenvectors. Compared to existing techniques, these new algorithms are at least an order of magnitude faster in high dimensions and drastically reduce memory requirements. The new algorithms provide instantaneous (few seconds) smoothing for matrices of dimension $J=10,000$ and very fast ($<$ 10 minutes) smoothing for $J=100,000$. Although SVDS is simpler than FACE, we provide ready to use, scalable R software for FACE. When incorporated into R package {it refund}, FACE improves the speed of penalized functional regression by an order of magnitude, even for data of normal size ($J<500$). We recommend that FACE be used in practice for the analysis of noisy and high-dimensional functional data.

Full work available at URL: https://arxiv.org/abs/1306.5718

Recommendations

• Fast covariance estimation for sparse functional data
• Fast symmetric additive covariance smoothing
• Fast BivariateP-Splines: The Sandwich Smoother
• Low-Rank Covariance Function Estimation for Multidimensional Functional Data
• Asymptotic efficiency in high-dimensional covariance estimation

zbMATH Keywords

smoothingsingular value decompositionfPCApenalized splinesFACEsandwich smoother

Mathematics Subject Classification ID

Nonparametric estimation (62G05) Factor analysis and principal components; correspondence analysis (62H25) Software, source code, etc. for problems pertaining to statistics (62-04) Estimation in multivariate analysis (62H12)

Cites Work

• Semiparametric Regression
• Multilevel functional principal component analysis
• Flexible smoothing with \(B\)-splines and penalties. With comments and a rejoinder by the authors
• Functional data analysis.
• Functional Data Analysis for Sparse Longitudinal Data
• Smoothing noisy data with spline functions: Estimating the correct degree of smoothing by the method of generalized cross-validation
• Thin Plate Regression Splines
• Simultaneous non-parametric regressions of unbalanced longitudinal data
• Applied functional data analysis. Methods and case studies
• Title not available (Why is that?)
• Title not available (Why is that?)
• Nonparametric Regression Analysis of Longitudinal Data
• Asymptotic theory for the principal component analysis of a vector random function: Some applications to statistical inference
• Principal components analysis of sampled functions
• Nonparametric estimation of common regressors for similar curve data
• Shrinkage estimation for functional principal component scores with application to the population kinetics of plasma folate
• Fast BivariateP-Splines: The Sandwich Smoother
• Statistical inferences for functional data
• Title not available (Why is that?)
• On the sample covariance matrix estimator of reduced effective rank population matrices, with applications to fPCA
• Generalized multilevel functional regression
• Smoothing Spline Gaussian Regression: More Scalable Computation via Efficient Approximation
• Confidence Intervals for Nonparametric Curve Estimates
• Nonparametric estimation of smoothed principal components analysis of sampled noisy functions
• An Accelerated-Time Model for Response Curves
• On the asymptotics of penalized spline smoothing
• Structured functional principal component analysis
• Longitudinal functional principal component analysis
• Longitudinal high-dimensional principal components analysis with application to diffusion tensor imaging of multiple sclerosis

Cited In (25)

• Functional analysis of spatial aggregation regions of Jeffrey pine Beetle-attack within the Lake Tahoe Basin
• Homogeneity tests of covariance for high-dimensional functional data with applications to event segmentation
• CovNet: Covariance Networks for Functional Data on Multidimensional Domains
• On variable ordination of Cholesky‐based estimation for a sparse covariance matrix
• Longitudinal high-dimensional principal components analysis with application to diffusion tensor imaging of multiple sclerosis
• Commentary: Methods for calculating growth trajectories and constructing growth centiles
• Bayesian inference and dynamic prediction of multivariate joint model with functional data: an application to Alzheimer's disease
• Fast Multilevel Functional Principal Component Analysis
• Dynamic child growth prediction: A comparative methods approach
• Comparing predictive abilities of longitudinal child growth models
• Scalable multiple changepoint detection for functional data sequences
• Optimal EMG placement for a robotic prosthesis controller with sequential, adaptive functional estimation (SAFE)
• Fast symmetric additive covariance smoothing
• Sensible functional linear discriminant analysis
• Gaussian graphical modeling for spectrometric data analysis
• Conditional analysis for mixed covariates, with application to feed intake of lactating sows
• Estimation of the nonparametric mean and covariance functions for multivariate longitudinal and sparse functional data
• Fixed-effects inference and tests of correlation for longitudinal functional data
• Fast covariance estimation for sparse functional data
• Linear scalar-on-surface random effects regression models
• Estimation for Extreme Conditional Quantiles of Functional Quantile Regression
• Additive Functional Cox Model
• Randomized estimation of functional covariance operator via subsampling
• Two-Dimensional Functional Principal Component Analysis for Image Feature Extraction
• Low-Rank Covariance Function Estimation for Multidimensional Functional Data

Uses Software

• R
• fda (R)
• SemiPar
• refund
• mgcv

This page was built for publication: Fast covariance estimation for high-dimensional functional data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2631375)