AI Paper Atlas

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

Leaky Hockey Stick Loss: The First Negatively Divergent Margin-based Loss Function for Classification

Oh-Ran Kwon, Hui Zou

JMLR 2023 · Learning theory and algorithms › Prediction and decision losses · approach: Ranking-based loss optimization

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
replace hinge loss with LHS lossKernel SVM · Simulation Ex2, n=200, Gaussian kernelmisclassification rate (%): 16.38 -> 15.45Improved
replace hinge loss with LHS lossKernel SVM · Simulation Ex2, n=900, linear kernelmisclassification rate (%): 38.11 -> 37.12Improved
replace hinge loss with LHS lossKernel SVM · Simulation Ex1, n=50, linear kernelmisclassification rate (%): 34.35 -> 34.63No clear effect
replace hinge loss with LHS lossKernel SVM · Simulation Ex3, n=50, Gaussian kernelmisclassification rate (%): 30.97 -> 31.36Hurt
replace hinge loss with LHS lossKernel SVM · UCI Sonar, linear kernelmisclassification rate (%): 25.10 -> 23.39Improved
replace hinge loss with LHS lossKernel SVM · UCI Banknote, Gaussian kernelmisclassification rate (%): 0.01 -> 0.44Hurt
replace hinge loss with LHS lossKernel SVM · UCI Musk, Gaussian kernelmisclassification rate (%): 8.29 -> 10.48Hurt
replace hinge loss with LHS lossKernel SVM · UCI datasets (15), linear kernelmisclassification rate (%): outperforms on ten data setsImproved
replace hinge loss with LHS lossKernel SVM · UCI datasets (15), Gaussian kernelmisclassification rate (%): outperforms on six data setsMixed
replace hinge loss with LHS lossKernel SVM · UCI Australian, linear kernel computation time (s)misclassification rate (%): 344.94 -> 0.35Improved
replace hinge loss with LHS lossKernel SVM · UCI Chess, Gaussian kernelmisclassification rate (%): 0.93 -> 1.70Hurt
increase sample size from 50 to 900LHS classifier · Simulation Ex1, Gaussian LHS classifiermisclassification rate (%): 17.11 -> 11.47Improved

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