Generalized Reference Kernel With Negative Samples For Support Vector One-class Classification

TL;DR

This paper introduces GRKneg, a novel kernel method for one-class SVMs that leverages negative samples to improve classification, especially with limited negative data, without requiring label information during optimization.

Contribution

The paper proposes a new kernel enhancement technique, GRKneg, that incorporates negative samples into one-class SVMs without using labels, improving performance with scarce negative data.

Findings

GRKneg outperforms standard OC-SVM with RBF kernel.

The method is especially effective with few negative samples.

Binary SVM performs better when many negative samples are available.

Abstract

This paper focuses on small-scale one-class classification with some negative samples available. We propose Generalized Reference Kernel with Negative Samples (GRKneg) for One-class Support Vector Machine (OC-SVM). We study different ways to select/generate the reference vectors and recommend an approach for the problem at hand. It is worth noting that the proposed method does not use any labels in the model optimization but uses the original OC-SVM implementation. Only the kernel used in the process is improved using the negative data. We compare our method with the standard OC-SVM and with the binary Support Vector Machine (SVM) using different amounts of negative samples. Our approach consistently outperforms the standard OC-SVM using Radial Basis Function kernel. When there are plenty of negative samples, the binary SVM outperforms the one-class approaches as expected, but we show that for the lowest numbers of negative samples the proposed approach clearly outperforms the binary SVM.