A partially linear framework for massive heterogeneous data

DOI10.1214/15-AOS1410zbMATH Open1358.62050arXiv1410.8570WikidataQ36352871 ScholiaQ36352871MaRDI QIDQ309709FDOQ309709

Han Liu, Tianqi Zhao, Guang Cheng

Publication date: 7 September 2016

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: We consider a partially linear framework for modelling massive heterogeneous data. The major goal is to extract common features across all sub-populations while exploring heterogeneity of each sub-population. In particular, we propose an aggregation type estimator for the commonality parameter that possesses the (non-asymptotic) minimax optimal bound and asymptotic distribution as if there were no heterogeneity. This oracular result holds when the number of sub-populations does not grow too fast. A plug-in estimator for the heterogeneity parameter is further constructed, and shown to possess the asymptotic distribution as if the commonality information were available. We also test the heterogeneity among a large number of sub-populations. All the above results require to regularize each sub-estimation as though it had the entire sample size. Our general theory applies to the divide-and-conquer approach that is often used to deal with massive homogeneous data. A technical by-product of this paper is the statistical inferences for the general kernel ridge regression. Thorough numerical results are also provided to back up our theory.

Full work available at URL: https://arxiv.org/abs/1410.8570

Recommendations

• Additive partially linear models for massive heterogeneous data
• Estimating population size of heterogeneous populations with large data sets and a large number of parameters
• Tuning-free heterogeneous inference in massive networks
• Partial least squares for heterogeneous data
• Projected spline estimation of the nonparametric function in high-dimensional partially linear models for massive data

zbMATH Keywords

heterogeneous datakernel ridge regressionpartially linear modeldivide-and-conquer methodmassive data

Mathematics Subject Classification ID

Point estimation (62F10) Asymptotic properties of parametric estimators (62F12) Asymptotic properties of nonparametric inference (62G20) Parametric tolerance and confidence regions (62F25) Ridge regression; shrinkage estimators (Lasso) (62J07)

Cites Work

• Semiparametric Regression
• Introduction to empirical processes and semiparametric inference
• Finite mixture models
• Smoothing Splines
• Title not available (Why is that?)
• Title not available (Why is that?)
• Additive regression and other nonparametric models
• On some global measures of the deviations of density function estimates
• Gaussian approximations and multiplier bootstrap for maxima of sums of high-dimensional random vectors
• An introduction to support vector machines and other kernel-based learning methods.
• Title not available (Why is that?)
• Title not available (Why is that?)
• \(\ell_{1}\)-penalization for mixture regression models
• Penalized quasi-likelihood estimation in partial linear models
• Statistical estimation in varying coefficient models
• Variable selection in semiparametric regression modeling
• Learning Bounds for Kernel Regression Using Effective Data Dimensionality
• The benefit of group sparsity
• On the asymptotic properties of the group lasso estimator for linear models
• Title not available (Why is that?)
• Semiparametric Regression for the Applied Econometrician
• PIECEWISE-POLYNOMIAL APPROXIMATIONS OF FUNCTIONS OF THE CLASSES $ W_{p}^{\alpha}$
• Sparse and efficient estimation for partial spline models with increasing dimension
• Joint asymptotics for semi-nonparametric regression models with partially linear structure
• Support union recovery in high-dimensional multivariate regression
• Local and global asymptotic inference in smoothing spline models
• Maximin effects in inhomogeneous large-scale data
• Divide and conquer kernel ridge regression: a distributed algorithm with minimax optimal rates
• Title not available (Why is that?)
• Title not available (Why is that?)
• A partially linear framework for massive heterogeneous data
• A split-and-conquer approach for analysis of
• A new concentration result for regularized risk minimizers
• Inference in Smoothing Spline Analysis of Variance
• Understanding Gaussian Process Regression Using the Equivalent Kernel
• New bounds on the Hermite polynomials

Cited In (71)

• A difference-based approach in the partially linear model with dependent errors
• Randomized sketches for kernel CCA
• Additive partially linear models for massive heterogeneous data
• Integrative Analysis for High-Dimensional Stratified Models
• Distributed subdata selection for big data via sampling-based approach
• A split-and-conquer variable selection approach for high-dimensional general semiparametric models with massive data
• Circumventing superefficiency: an effective strategy for distributed computing in non-standard problems
• Projected spline estimation of the nonparametric function in high-dimensional partially linear models for massive data
• First-Order Newton-Type Estimator for Distributed Estimation and Inference
• Model aggregation for doubly divided data with large size and large dimension
• Distributed estimation with empirical likelihood
• High-dimensional integrative analysis with homogeneity and sparsity recovery
• Distributed testing and estimation under sparse high dimensional models
• Title not available (Why is that?)
• Communication-efficient estimation of high-dimensional quantile regression
• High-Dimensional Factor Regression for Heterogeneous Subpopulations
• Optimal subsample selection for massive logistic regression with distributed data
• Distributed estimation in heterogeneous reduced rank regression: with application to order determination in sufficient dimension reduction
• Robust reduced rank regression in a distributed setting
• Distributed Estimation for Principal Component Analysis: An Enlarged Eigenspace Analysis
• A partially linear framework for massive heterogeneous data
• Title not available (Why is that?)
• Quantile regression in big data: a divide and conquer based strategy
• Divide and conquer in nonstandard problems and the super-efficiency phenomenon
• Smoothing Spline Semiparametric Density Models
• Automatic variable selection in a linear model on massive data
• Distributed linear regression by averaging
• WONDER: Weighted one-shot distributed ridge regression in high dimensions
• Title not available (Why is that?)
• Distributed Sparse Composite Quantile Regression in Ultrahigh Dimensions
• A Massive Data Framework for M-Estimators with Cubic-Rate
• Quantile regression under memory constraint
• An adaptive lack of fit test for big data
• An RKHS-based approach to double-penalized regression in high-dimensional partially linear models
• Feasible algorithm for linear mixed model for massive data
• Linear operator‐based statistical analysis: A useful paradigm for big data
• Title not available (Why is that?)
• Title not available (Why is that?)
• Distributed Generalized Cross-Validation for Divide-and-Conquer Kernel Ridge Regression and Its Asymptotic Optimality
• Distributed adaptive Huber regression
• Off-Policy Estimation of Long-Term Average Outcomes With Applications to Mobile Health
• Efficient fused learning for distributed imbalanced data
• Partitioned Approach for High-dimensional Confidence Intervals with Large Split Sizes
• Distributed Bayesian posterior voting strategy for massive data
• A review of distributed statistical inference
• Title not available (Why is that?)
• Score-matching representative approach for big data analysis with generalized linear models
• Distributed optimization and statistical learning for large-scale penalized expectile regression
• Nonparametric Bayesian Aggregation for Massive Data
• Title not available (Why is that?)
• A scalable nonparametric specification testing for massive data
• Computational Limits of A Distributed Algorithm For Smoothing Spline
• Approximating Partial Likelihood Estimators via Optimal Subsampling
• Distributed debiased estimation of high-dimensional partially linear models with jumps
• A communication efficient distributed one-step estimation
• Distributed statistical optimization for non-randomly stored big data with application to penalized learning
• Aggregated inference
• Distributed estimation and inference for semiparametric binary response models
• Distributed estimation of functional linear regression with functional responses
• Statistical Inference on Panel Data Models: A Kernel Ridge Regression Method
• Divide-and-Conquer: A Distributed Hierarchical Factor Approach to Modeling Large-Scale Time Series Data
• Global debiased DC estimations for biased estimators via pro forma regression
• DDAC-SpAM: A Distributed Algorithm for Fitting High-dimensional Sparse Additive Models with Feature Division and Decorrelation
• Online two-way estimation and inference via linear mixed-effects models
• Statistical Inference for Maximin Effects: Identifying Stable Associations across Multiple Studies
• Support vector machine in big data: smoothing strategy and adaptive distributed inference
• Adaptive distributed inference for multi-source massive heterogeneous data
• Title not available (Why is that?)
• Sampling-based estimation for massive survival data with additive hazards model
• Robust distributed multicategory angle-based classification for massive data
• Statistical inference and distributed implementation for linear multicategory SVM

Uses Software

• gss
• SemiPar

This page was built for publication: A partially linear framework for massive heterogeneous data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q309709)