Abstract:Support Vector Machines (SVMs) were primarily designed for 2-class classification. But they have been extended for N-class classification also based on the requirement of multiclasses in the practical applications. Although N-class classification using SVM has considerable research attention, getting minimum number of classifiers at the time of training and testing is still a continuing research. We propose a new algorithm CBTS-SVM (Centroid based Binary Tree Structured SVM) which addresses this issue. In this we build a binary tree of SVM models based on the similarity of the class labels by finding their distance from the corresponding centroids at the root level. The experimental results demonstrates the comparable accuracy for CBTS with OVO with reasonable gamma and cost values. On the other hand when CBTS is compared with OVA, it gives the better accuracy with reduced training time and testing time. Furthermore CBTS is also scalable as it is able to handle the large data sets.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to reduce the number of binary classifiers required during training and testing in multi - class Support Vector Machines (SVMs), while maintaining high classification accuracy. Specifically: 1. **Background problems**: - Support Vector Machines (SVMs) were originally designed for binary classification problems. - In practical applications, many scenarios require dealing with multi - class problems (for example, classifying astronomical objects as stars, galaxies or quasars). 2. **Limitations of existing methods**: - One method is to decompose multi - class problems into multiple binary classification problems, such as One - vs - All (OVA) and One - vs - One (OVO). - These methods require a large number of binary classifiers during training and testing, resulting in high computational time and resource consumption. 3. **New method proposed in the paper**: - A Centroid - Based Binary Tree Structure SVM (CBTS - SVM) is proposed, which reduces the number of required binary classifiers by constructing a binary tree. - Use K - Means clustering as a pre - processing step to divide the data into two disjoint sets according to the similarity between class labels, and recursively construct nodes in the binary tree. 4. **Objectives**: - Reduce training and testing time. - Improve classification accuracy, especially when dealing with large - scale data sets. 5. **Experimental results**: - The experimental results show that CBTS - SVM is superior to OVA in training and testing time, and is comparable or even better than OVO in accuracy. - CBTS - SVM can handle large - scale data sets and has good scalability. In summary, this paper aims to solve the problem of low training and testing efficiency in multi - class SVMs by introducing the CBTS - SVM algorithm, while maintaining high classification accuracy.

Centroid Based Binary Tree Structured SVM for Multi Classification

A Novel Svm Modeling Approach For Highly Imbalanced And Overlapping Classification

A novel tree kernel support vector machine classifier for modeling the relationship between bioactivity and molecular descriptors

A Multi-Class Large Margin Classifier

Novel One-Vs-rest Classifier Based on SVM and Multi-Spheres

Improved support vector machine for multi-class classification problems

Multivariate time series classification based on μσ-DWC feature and tree-structured M-SVM

Novel Design of Decision-Tree-Based Support Vector Machines Multi-Class Classifier

DCSVM: fast multi-class classification using support vector machines

Multi-sphere Classification Algorithm Based on 1-SVM

Nonparallel Hyperplane Classifiers for Multi-category Classification

Multiclass Approaches for Support Vector Machine Based Land Cover Classification

Enhancing multiclass brain tumor diagnosis using SVM and innovative feature extraction techniques

Decision Tree Support Vector Machine Based on Genetic Algorithm for Multi-Class Classification

Hierarchical 1-FSVM Algorithm Based on SMO

Rough Set Based Multi-Class Core Vector Machine

Multi-class fault diagnosis based on support vector machines with sequenced binary tree archtecture

Multi-Class SVM Classifier Based on Pairwise Coupling

Hyperspectral Remote Sensing Image Classification Based on Support Vector Machine

Studies on the evaluation of college classroom teaching quality based on SVM multiclass classification algorithm

Unfolding multilevel agglomerative strategies for SVM classification: a case study in discriminating spectrally similar land covers