GBSVR: Granular Ball Support Vector Regression

Reshma Rastogi, Ankush Bisht, Sanjay Kumar, Suresh Chandra · 2025

Read GBSVR: Granular Ball Support Vector Regression by Reshma Rastogi, Ankush Bisht, Sanjay Kumar, Suresh Chandra (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-10539.txt.

Description

Support vector regression is a sturdy prediction tool that turns slow on big data, since its training solves a heavy optimization problem. Reshma Rastogi, Ankush Bisht, Sanjay Kumar, and Suresh Chandra speed it up in a 2025 paper by first bundling the data. Their granular balls group nearby points, so the solver sees a handful of balls instead of millions of points.

The bundling also dulls the sting of outliers, a known weakness of the method's loss function. A discretization step for continuous inputs helps the balls form. On standard benchmark sets, the bundled version beat the leading rival methods the team tested. The authors present granular ball regression as a way to keep the support vector family's accuracy while shedding the cost that kept it off large datasets.

Catalog record

CREATORReshma Rastogi, Ankush Bisht, Sanjay Kumar, Suresh Chandra
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-10539

Browse the full library · About the Tower of Babel