Fast algorithms for least square problems with Kronecker lower subsets

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Malik, Osman Asif, Xu, Yiming, Cheng, Nuojin, Becker, Stephen, Doostan, Alireza, Narayan, Akil
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910919153418240
author Malik, Osman Asif
Xu, Yiming
Cheng, Nuojin
Becker, Stephen
Doostan, Alireza
Narayan, Akil
author_facet Malik, Osman Asif
Xu, Yiming
Cheng, Nuojin
Becker, Stephen
Doostan, Alireza
Narayan, Akil
contents While leverage score sampling provides powerful tools for approximating solutions to large least squares problems, the cost of computing exact scores and sampling often prohibits practical application. This paper addresses this challenge by developing a new and efficient algorithm for exact leverage score sampling applicable to matrices that are lower column subsets of Kronecker product matrices. We synthesize relevant approximation guarantees and detail the algorithm that specifically leverages this structural property for computational efficiency. Through numerical examples, we demonstrate that utilizing efficiently computed exact leverage scores via our methods significantly reduces approximation errors, as compared to established approximate leverage score sampling strategies when applied to this important class of structured matrices.
format Preprint
id arxiv_https___arxiv_org_abs_2209_05662
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Fast algorithms for least square problems with Kronecker lower subsets
Malik, Osman Asif
Xu, Yiming
Cheng, Nuojin
Becker, Stephen
Doostan, Alireza
Narayan, Akil
Numerical Analysis
Data Structures and Algorithms
While leverage score sampling provides powerful tools for approximating solutions to large least squares problems, the cost of computing exact scores and sampling often prohibits practical application. This paper addresses this challenge by developing a new and efficient algorithm for exact leverage score sampling applicable to matrices that are lower column subsets of Kronecker product matrices. We synthesize relevant approximation guarantees and detail the algorithm that specifically leverages this structural property for computational efficiency. Through numerical examples, we demonstrate that utilizing efficiently computed exact leverage scores via our methods significantly reduces approximation errors, as compared to established approximate leverage score sampling strategies when applied to this important class of structured matrices.
title Fast algorithms for least square problems with Kronecker lower subsets
topic Numerical Analysis
Data Structures and Algorithms
url https://arxiv.org/abs/2209.05662