Reducing number of classifiers in DAGSVM based on class similarity

Authors:
Marcin Luckner
Affiliations:
Warsaw University of Technology, Faculty of Mathematics and Information Science, Warsaw, Poland
Venue:
ICIAP'11 Proceedings of the 16th international conference on Image analysis and processing: Part I
Year:
2011

Citing 9
Cited 0

Introduction to graph theory

Introduction to graph theory
The nature of statistical learning theory

The nature of statistical learning theory
Pairwise classification and support vector machines

Advances in kernel methods
Combining support vector and mathematical programming methods for classification

Advances in kernel methods
Adaptive Directed Acyclic Graphs for Multiclass Classification

PRICAI '02 Proceedings of the 7th Pacific Rim International Conference on Artificial Intelligence: Trends in Artificial Intelligence
On the Learnability and Design of Output Codes for Multiclass Problems

COLT '00 Proceedings of the Thirteenth Annual Conference on Computational Learning Theory
Support Vector Machines for Pattern Classification (Advances in Pattern Recognition)

Support Vector Machines for Pattern Classification (Advances in Pattern Recognition)
Reducing the number of sub-classifiers for pairwise multi-category support vector machines

Pattern Recognition Letters
A comparison study on multiple binary-class SVM methods for unilabel text categorization

Pattern Recognition Letters

Quantified Score

Hi-index	0.01

Visualization

Abstract

Support Vector Machines are excellent binary classifiers. In case of multi-class classification problems individual classifiers can be collected into a directed acyclic graph structure DAGSVM. Such structure implements One-Against-One strategy. In this strategy a split is created for each pair of classes, but, because of hierarchical structure, only a part of them is used in the single classification process. The number of classifiers may be reduced if their classification tasks will be changed from separation of individual classes into separation of groups of classes. The proposed method is based on the similarity of classes. For near classes the structure of DAG stays immutable. For the distant classes more than one is separated with a single classifier. This solution reduces the classification cost. At the same time the recognition accuracy is not reduced in a significant way. Moreover, a number of SV, which influences on the learning time will not grow rapidly.