Abstract:This paper addresses the challenge of integrating sequentially arriving data within the quantile regression framework, where the number of features is allowed to grow with the number of observations, the horizon is unknown, and memory is limited. We employ stochastic sub-gradient descent to minimize the empirical check loss and study its statistical properties and regret performance. In our analysis, we unveil the delicate interplay between updating iterates based on individual observations versus batches of observations, revealing distinct regularity properties in each scenario. Our method ensures long-term optimal estimation irrespective of the chosen update strategy. Importantly, our contributions go beyond prior works by achieving exponential-type concentration inequalities and attaining optimal regret and error rates that exhibit only \textsf{ short-term} sensitivity to initial errors. A key insight from our study is the delicate statistical analyses and the revelation that appropriate stepsize schemes significantly mitigate the impact of initial errors on subsequent errors and regrets. This underscores the robustness of stochastic sub-gradient descent in handling initial uncertainties, emphasizing its efficacy in scenarios where the sequential arrival of data introduces uncertainties regarding both the horizon and the total number of observations. Additionally, when the initial error rate is well-controlled, there is a trade-off between short-term error rate and long-term optimality. Due to the lack of delicate statistical analysis for squared loss, we also briefly discuss its properties and proper schemes. Extensive simulations support our theoretical findings.

Quantile Regression With L(1)-Regularization And Gaussian Kernels

Quantile Inference for Moderate Deviations from a Unit Root Model with Infinite Variance

Kernel Regression with Coefficient-based l(q)-regularization

Regularization Strategies for Quantile Regression

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

Sparse Kernel Regression with Coefficient-based $\Ell_q-$Regularization.

Structured kernel quantile regression

Bayesian Regularized Quantile Regression

Sparse quantile regression

Concentration Estimates for Learning with ℓ1-Regularizer and Data Dependent Hypothesis Spaces

Approximate nonparametric quantile regression in reproducing kernel Hilbert spaces via random projection

Online Quantile Regression

Linear Convergence of Inexact Descent Method and Inexact Proximal Gradient Algorithms for Lower-Order Regularization Problems

Optimal Prediction of Quantile Functional Linear Regression in Reproducing Kernel Hilbert Spaces

Optimal Kernel Quantile Learning with Random Features

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

Partially linear functional quantile regression in a reproducing kernel Hilbert space

Small noise analysis for Tikhonov and RKHS regularizations

Analysis of Least Squares Regularized Regression in Reproducing Kernel Krein Spaces

Statistical Performance of Quantile Tensor Regression with Convex Regularization

Kernel Type Smoothed Quantile Estimation under Long Memory