Multi-class nonparallel support vector machine

In this paper, we propose an extended version of Nonparallel Support Vector Machine (NPSVM) for multi-classification using one-versus-one-versus-rest approach called MCNPSVM. The MCNPSVM is converted into a series of binary classification problems, in each of which two nonparallel hyperplanes are fo...

Full description

Saved in:
Bibliographic Details
Published in:Progress in artificial intelligence 2023-12, Vol.12 (4), p.349-361
Main Authors: Sahleh, Ali, Salahi, Maziar, Eskandari, Sadegh
Format: Article
Language:English
Subjects:
Artificial Intelligence
Classification
Computational Intelligence
Computer Imaging
Computer Science
Control
Data Mining and Knowledge Discovery
Data points
Hyperplanes
Mechatronics
Natural Language Processing (NLP)
Pattern Recognition and Graphics
Quadratic programming
Regular Paper
Robotics
Support vector machines
Vision
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we propose an extended version of Nonparallel Support Vector Machine (NPSVM) for multi-classification using one-versus-one-versus-rest approach called MCNPSVM. The MCNPSVM is converted into a series of binary classification problems, in each of which two nonparallel hyperplanes are found by solving two quadratic programming problems. This is done in such a way that each hyperplane is aligned with the data points of the class that it represents by constructing an ϵ -insensitive tube and is as far as possible from the other class, while the rest of the data are in the margin of these two hyperplanes. Further, in order to accelerate learning time of MCNPSVM, the mean of data matrix corresponding to the rest classes is used. Experiments on benchmark datasets are executed to study the performance of proposed models compared to various multi-class SVM extensions.
ISSN:2192-6352
2192-6360
DOI:10.1007/s13748-023-00308-7