AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Support Vector Machine Classifier via Soft-Margin Loss

Huajun Wang, Yuan‐Hai Shao, Shenglong Zhou, Ce Zhang, Naihua Xiu

TPAMI 2021 · Learning theory and algorithms › Prediction and decision losses · approach: Ranking-based loss optimization · 169 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use L0/1 soft-margin loss with ADMM instead of hinge loss (LibSVM)SVM classifier · Australian (real data)number of support vectors: 203 -> 24Improved
use L0/1 soft-margin loss instead of hinge lossSVM classifier · A6a (real data)number of support vectors: 4346 -> 370Improved
use L0/1 soft-margin loss instead of hinge lossSVM classifier · Colon-cancer (real data)test accuracy (%): 64.52 -> 90.23Improved
use L0/1 soft-margin loss instead of hinge lossSVM classifier · Splice (real data)test accuracy (%): 88.97 -> 85.52Hurt
use L0/1 soft-margin loss instead of Katyusha-trained squared hingeSVM classifier · HIGGS (real data)test accuracy (%): 59.46 -> 65.21Improved
use L0/1 soft-margin loss instead of hinge lossSVM classifier · Synthetic 2D without outliers (m=10000)CPU time (s): 0.094 -> 0.018Improved
use L0/1 soft-margin loss instead of hinge lossSVM classifier · Synthetic 2D without outliers (m=10000)test accuracy (%): 97.18 -> 97.20No clear effect
use L0/1 soft-margin loss instead of hinge loss under 20% label flipsSVM classifier · Synthetic 2D with outliers (m=5000)test accuracy (%): 78.16 -> 78.30Improved
use L0/1 soft-margin loss instead of hinge loss under 20% label flipsSVM classifier · Synthetic 2D with outliers (m=5000)number of support vectors: 2160 -> 13Improved
use L0/1 loss on very large data (HIGGS, 11M samples)SVM classifier · HIGGS (real data)solver feasibilityImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy