Abstract:Standard regression techniques model only the mean of the response variable. Quantile regression (QR) is more powerful in that it depicts a comprehensive relationship between the response variable and independent covariates at different quantiles. It is particularly useful for non-normally distributed data with skewness or heterogeneity, which appear routinely in many scientific fields, such as economics, finance, public health and biology. Although its theory has been well developed in the literature, its computation in big data still faces multiple challenges, especially for vertically stored big data in modern distributed environments, where communication efficiency and security are usually the primary considerations. While the popular alternating direction method of multipliers (ADMM) provides a general computational solution, its slow convergence becomes a bottleneck when communication cost dominates local computational consumption, such as Internet of Things (IoT) networks. Motivated by the residual projection technique, in this paper we propose an innovative iterative parallel framework, PIQR, that converges faster and has a more secure data transmission plan, and establish its convergence property. This framework is further extended to composite quantile regression (CQR), a modified QR technique that improves estimation efficiency at extreme quantiles. Simulation studies show that both the ADMM-based method and the PIQR enjoy favorable estimation accuracy in distributed environments. While PIQR is inferior to the ADMM-based method at local computation, it requires much fewer iterations to achieve convergence, and hence significantly improves the overall computational efficiency when communication cost is the dominating factor. Moreover, PIQR transmits only data involving the residual information between different machines, and can better prevent the leakage of important data information compared with the ADMM-based method.

Communication-efficient Estimation of Quantile Matrix Regression for Massive Datasets

Estimation of Heteroscedasticity by Local Composite Quantile Regression and Matrix Decomposition

Distributed Bootstrap Simultaneous Inference for High-Dimensional Quantile Regression

Distributed quantile regression for massive heterogeneous data

Distributed Online Quantile Regression over Networks with Quantized Communication.

Distributed Quantile Regression over Sensor Networks

Sparse and Low-Rank Matrix Quantile Estimation with Application to Quadratic Regression

Communication-Efficient Nonparametric Quantile Regression via Random Features

Communication-efficient estimation and inference for high-dimensional quantile regression based on smoothed decorrelated score

Distributed High-Dimensional Quantile Regression: Estimation Efficiency and Support Recovery

Fast and Robust Low-Rank Learning over Networks: A Decentralized Matrix Quantile Regression Approach

Sparse Estimation Via ℓ_q Optimization Method in High-Dimensional Linear Regression

Distributed High-dimensional Regression under a Quantile Loss Function

A Note on Distributed Quantile Regression by Pilot Sampling and One-Step Updating

Online Quantile Regression

Regularized Adaptive Huber Matrix Regression and Distributed Learning

Distributed Censored Regression over Networks

Residual projection for quantile regression in vertically partitioned big data

Low-rank matrix estimation in multi-response regression with measurement errors: Statistical and computational guarantees.

Distributed Sparse Precision Matrix Estimation via Alternating Block-Based Gradient Descent

Distributed optimization and statistical learning for large-scale penalized expectile regression