Saved in:
Bibliographic Details
Main Authors: Tazik, Ladan, Stafford, James, Braun, John
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.18695
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916707869655040
author Tazik, Ladan
Stafford, James
Braun, John
author_facet Tazik, Ladan
Stafford, James
Braun, John
contents The local least squares estimator for a regression curve cannot provide optimal results when non-Gaussian noise is present. Both theoretical and empirical evidence suggests that residuals often exhibit distributional properties different from those of a normal distribution, making it worthwhile to consider estimation based on other norms. It is suggested that $L_p$-norm estimators be used to minimize the residuals when these exhibit non-normal kurtosis. In this paper, we propose a local polynomial $L_p$-norm regression that replaces weighted least squares estimation with weighted $L_p$-norm estimation for fitting the polynomial locally. We also introduce a new method for estimating the parameter $p$ from the residuals, enhancing the adaptability of the approach. Through numerical and theoretical investigation, we demonstrate our method's superiority over local least squares in one-dimensional data and show promising outcomes for higher dimensions, specifically in 2D.
format Preprint
id arxiv_https___arxiv_org_abs_2504_18695
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Local Polynomial Lp-norm Regression
Tazik, Ladan
Stafford, James
Braun, John
Machine Learning
Other Statistics
The local least squares estimator for a regression curve cannot provide optimal results when non-Gaussian noise is present. Both theoretical and empirical evidence suggests that residuals often exhibit distributional properties different from those of a normal distribution, making it worthwhile to consider estimation based on other norms. It is suggested that $L_p$-norm estimators be used to minimize the residuals when these exhibit non-normal kurtosis. In this paper, we propose a local polynomial $L_p$-norm regression that replaces weighted least squares estimation with weighted $L_p$-norm estimation for fitting the polynomial locally. We also introduce a new method for estimating the parameter $p$ from the residuals, enhancing the adaptability of the approach. Through numerical and theoretical investigation, we demonstrate our method's superiority over local least squares in one-dimensional data and show promising outcomes for higher dimensions, specifically in 2D.
title Local Polynomial Lp-norm Regression
topic Machine Learning
Other Statistics
url https://arxiv.org/abs/2504.18695