Learning rates for kernel-based expectile regression

DOI10.1007/S10994-018-5762-9zbMATH Open1480.62067arXiv1702.07552OpenAlexW2592603639WikidataQ129210583 ScholiaQ129210583MaRDI QIDQ669274FDOQ669274

Authors: Ingo Steinwart, M. U. Farooq 

Publication date: 15 March 2019

Published in: Machine Learning (Search for Journal in Brave)

Abstract: Conditional expectiles are becoming an increasingly important tool in finance as well as in other areas of applications. We analyse a support vector machine type approach for estimating conditional expectiles and establish learning rates that are minimax optimal modulo a logarithmic factor if Gaussian RBF kernels are used and the desired expectile is smooth in a Besov sense. As a special case, our learning rates improve the best known rates for kernel-based least squares regression in this scenario. Key ingredients of our statistical analysis are a general calibration inequality for the asymmetric least squares loss, a corresponding variance bound as well as an improved entropy number bound for Gaussian RBF kernels.

Full work available at URL: https://arxiv.org/abs/1702.07552

Recommendations

• An SVM-like approach for expectile regression
• Optimal regression rates for SVMs using Gaussian kernels
• KLERC: kernel Lagrangian expectile regression calculator
• Conditional quantiles with varying Gaussians
• Learning rates for the risk of kernel-based quantile regression estimators in additive models

zbMATH Keywords

support vector machinesvariance boundlearning ratesentropy number boundself-calibration inequality

Mathematics Subject Classification ID

Nonparametric regression and quantile regression (62G08) Applications of statistics to economics (62P20) Learning and adaptive systems in artificial intelligence (68T05) Limit theorems for vector-valued random variables (infinite-dimensional case) (60B12)

Cites Work

• Title not available (Why is that?)
• Regression Quantiles
• Title not available (Why is that?)
• Asymmetric Least Squares Estimation and Testing
• Support Vector Machines
• Adaptive Bayesian estimation using a Gaussian random field with inverse gamma bandwidth
• On the mathematical foundations of learning
• Relating quantiles and expectiles under weighted-symmetry
• Generalized quantiles as risk measures
• Coherence and elicitability
• Geoadditive expectile regression
• Asymmetric least squares regression estimation: A nonparametric approach^∗
• Nonparametric multiple expectile regression via ER-Boost
• Title not available (Why is that?)
• A distribution-free theory of nonparametric regression
• Title not available (Why is that?)
• Optimal rates for the regularized least-squares algorithm
• Support vector machine soft margin classifiers: error analysis
• Title not available (Why is that?)
• Interpolation of Besov Spaces
• Besov Spaces on Domains in ℝ d
• Learning rates of least-square regularized regression
• Estimating conditional quantiles with the help of the pinball loss
• Statistical performance of support vector machines
• A new concentration result for regularized risk minimizers
• Model selection for regularized least-squares algorithm in learning theory
• On regularization algorithms in learning theory
• An introduction to Sobolev spaces and interpolation spaces
• Regularization in kernel learning
• Training SVMs without offset
• Optimal regression rates for SVMs using Gaussian kernels
• How to compare different loss functions and their risks
• Maximum-gain working set selection for SVMs
• Title not available (Why is that?)
• Oracle inequalities for support vector machines that are based on random entropy numbers
• An SVM-like approach for expectile regression
• Nonlinear expectile regression with application to value-at-risk and expected shortfall estimation
• A dynamic autoregressive expectile for time-invariant portfolio protection strategies
• Title not available (Why is that?)
• Nonparametric regression expectiles^∗
• Nonparametric conditional autoregressive expectile model via neural network with applications to estimating financial risk

Cited In (13)

• Nonparametric estimation of expectile regression in functional dependent data
• Local linear estimate of the functional expectile regression
• Optimal learning rates for kernel partial least squares
• A closer look at covering number bounds for Gaussian kernels
• Learning rates for the kernel regularized regression with a differentiable strongly convex loss
• Title not available (Why is that?)
• Connection between higher order measures of risk and stochastic dominance
• Analysis of regularized least-squares in reproducing kernel Kreĭn spaces
• The local linear functional \(k\)NN estimator of the conditional expectile: uniform consistency in number of neighbors
• Kernel estimation under linear-exponential loss
• Adaptive learning rates for support vector machines working on data with low intrinsic dimension
• The functional \(k\mathrm{NN}\) estimator of the conditional expectile: uniform consistency in number of neighbors
• Optimal learning with Gaussians and correntropy loss

Uses Software

• GURLS
• erboost

This page was built for publication: Learning rates for kernel-based expectile regression

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q669274)