Functional principal subspace sampling for large scale functional data analysis

DOI10.1214/22-EJS2010zbMATH Open1493.62642arXiv2109.03397MaRDI QIDQ2137809FDOQ2137809

Authors: Shiyuan He, Xiaomeng Yan 

Publication date: 11 May 2022

Published in: Electronic Journal of Statistics (Search for Journal in Brave)

Abstract: Functional data analysis (FDA) methods have computational and theoretical appeals for some high dimensional data, but lack the scalability to modern large sample datasets. To tackle the challenge, we develop randomized algorithms for two important FDA methods: functional principal component analysis (FPCA) and functional linear regression (FLR) with scalar response. The two methods are connected as they both rely on the accurate estimation of functional principal subspace. The proposed algorithms draw subsamples from the large dataset at hand and apply FPCA or FLR over the subsamples to reduce the computational cost. To effectively preserve subspace information in the subsamples, we propose a functional principal subspace sampling probability, which removes the eigenvalue scale effect inside the functional principal subspace and properly weights the residual. Based on the operator perturbation analysis, we show the proposed probability has precise control over the first order error of the subspace projection operator and can be interpreted as an importance sampling for functional subspace estimation. Moreover, concentration bounds for the proposed algorithms are established to reflect the low intrinsic dimension nature of functional data in an infinite dimensional space. The effectiveness of the proposed algorithms is demonstrated upon synthetic and real datasets.

Full work available at URL: https://arxiv.org/abs/2109.03397

Recommendations

• Multi-dimensional functional principal component analysis
• Analysing large datasets of functional data: a survey sampling point of view
• Sampled forms of functional PCA in reproducing kernel Hilbert spaces
• Functional principal components analysis via penalized rank one approximation
• An algorithm for the principal component analysis of large data sets

zbMATH Keywords

randomized algorithmfunctional principal component analysisfunctional linear regressionoperator perturbation theory

Mathematics Subject Classification ID

Factor analysis and principal components; correspondence analysis (62H25) Functional data analysis (62R10) Statistical sampling theory and related topics (62D99)

Cites Work

• Inference for functional data with applications
• Principal component models for sparse functional data
• Functional Data Analysis for Sparse Longitudinal Data
• Introduction to Functional Data Analysis
• Theoretical foundations of functional data analysis, with an introduction to linear operators
• Finding structure with randomness: probabilistic algorithms for constructing approximate matrix decompositions
• Functional linear regression analysis for longitudinal data
• Defining probability density for a distribution of random functions
• Fast Monte Carlo Algorithms for Matrices III: Computing a Compressed Approximate Matrix Decomposition
• An introduction to matrix concentration inequalities
• A statistical perspective on algorithmic leveraging
• Faster least squares approximation
• A reproducing kernel Hilbert space approach to functional linear regression
• Methodology and convergence rates for functional linear regression
• Fast Monte Carlo Algorithms for Matrices II: Computing a Low-Rank Approximation to a Matrix
• CLT in functional linear regression models
• Estimation of the Mean of Functional Time Series and a Two-Sample Problem
• Fast Monte Carlo Algorithms for Matrices I: Approximating Matrix Multiplication
• Horvitz-Thompson estimators for functional data: asymptotic confidence bands and optimal allocation for stratified sampling
• Randomized Sketches of Convex Programs With Sharp Guarantees
• Fast approximation of matrix coherence and statistical leverage
• Asymptotics and concentration bounds for bilinear forms of spectral projectors of sample covariance
• Kernel ridge vs. principal component regression: minimax bounds and the qualification of regularization operators
• On some extensions of Bernstein's inequality for self-adjoint operators
• Estimation in functional linear quantile regression
• Properties of design-based functional principal components analysis
• Randomized sketches for kernels: fast and optimal nonparametric regression
• Sketching as a tool for numerical linear algebra
• Optimal subsampling for large sample logistic regression
• More efficient estimation for logistic regression with optimal subsamples
• Subspace Sampling and Relative-Error Matrix Approximation: Column-Row-Based Methods
• A statistical perspective on randomized sketching for ordinary least-squares
• Sketched ridge regression: optimization perspective, statistical perspective, and model averaging
• Rotation sampling for functional data
• Uniform convergence and asymptotic confidence bands for model-assisted estimators of the mean of sampled functional data
• Newton Sketch: A Near Linear-Time Optimization Algorithm with Linear-Quadratic Convergence
• Confidence bands for Horvitz-Thompson estimators using sampled noisy functional data
• Sketching for principal component regression
• Optimal Sampling for Generalized Linear Models Under Measurement Constraints
• Lectures on randomized numerical linear algebra
• Analysing large datasets of functional data: a survey sampling point of view
• OSUMC

Cited In (3)

• Smart sampling and incremental function learning for very large high dimensional data
• Optimal subsampling for functional quantile regression
• Functional Horseshoe Priors for Subspace Shrinkage

Uses Software

• OSUMC

This page was built for publication: Functional principal subspace sampling for large scale functional data analysis

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2137809)