Abstract:Theoretically, support vector machines (SVMs) have general error bounds along a radius-margin ratio, while conventional SVMs consider only the maximization of the margin and ignore the minimization of the radius, which is sensitive to affine data transformations. Thus, conventional SVMs can be improved by controlling both the radius and the margin. Several SVM variants based on radius-margin ratio error bounds have been proposed to integrate the radius and margin. However, most of these either require a diagonal transformation matrix or are computationally expensive to optimize. In this paper, we propose a novel radius-margin-based SVM model with LogDet regularization called L-SVM. In our model, we consider the radius and introduce a negative LogDet term to improve the model accuracy. We also adopt a two-step alternating minimization strategy to obtain an optimal solution, which leads to impressive computational improvements. Our experimental results validate the performance of the L-SVM and show that the L-SVM achieves significantly higher accuracy and efficiency compared to conventional SVMs and some other state-of-the-art radius-margin-based SVM methods. In addition, we apply our proposed L-SVM to solve transaction fraud problems and propose a framework for an L-SVM-based fraud detection system. (C) 2018 Elsevier Ltd. All rights reserved.

Theory and Analysis of Double-Margin SVM

A Multi-Class Large Margin Classifier

An Unbalanced Dataset Classification Approach Based On V-Support Vector Machine

A novel twin large margin distribution machine

Multi-label Optimal Margin Distribution Machine

Margin Distribution Analysis

Mining Knowledge from Unbalanced Data: Effect of Class Distribution on SVM Classification

A Double Margin Based Fuzzy Support Vector Machine Algorithm

On the Optimization of Margin Distribution

μSVM - A new method for solving the problem of imbalanced dataset classification

Large Margin Kernel Pocket Algorithm

A Rough Margin Based Support Vector Machine.

Recent Advances in Large Margin Learning

Research and Analysis of Methods for Multiclass Support Vector Machines

Analysis of Classification Margin for Classification Accuracy with Applications

Structured Large Margin Machine Ensemble

On Ψ-Learning

On Psi-Learning

Scalable Optimal Margin Distribution Machine

L-SVM: A Radius-Margin-based SVM Algorithm with LogDet Regularization.

Large Margin Deep Neural Networks: Theory and Algorithms.