Multiple Sources Data Fusion Strategies Based on Multi-class Support Vector Machine

Authors:
Luo Zhong;Zhe Li;Zichun Ding;Cuicui Guo;Huazhu Song
Affiliations:
School of Computer Science and Technology, Wuhan University of Technology, Wuhan, P.R.C 430070;School of Computer Science and Technology, Wuhan University of Technology, Wuhan, P.R.C 430070;School of Computer Science and Technology, Wuhan University of Technology, Wuhan, P.R.C 430070;School of Computer Science and Technology, Wuhan University of Technology, Wuhan, P.R.C 430070;School of Computer Science and Technology, Wuhan University of Technology, Wuhan, P.R.C 430070
Venue:
ISNN '08 Proceedings of the 5th international symposium on Neural Networks: Advances in Neural Networks
Year:
2008

Citing 3
Cited 0

The nature of statistical learning theory

The nature of statistical learning theory
Probability Estimates for Multi-class Classification by Pairwise Coupling

The Journal of Machine Learning Research
A comparison of methods for multiclass support vector machines

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Data fusion from multiple sources is very important and useful. Some strategies for multiple sources data fusion based on Multi-class Support Vector Machine (MSVM) are proposed in this paper. The features (independent components) of data from multiple sources are extracted for fusion. The Dempster-Shafer theory (DS theory) and Bayesian theory are used to combine the probabilistic outputs of MSVMs. Then the outputs of DS theory combination are classified by a MSVM to get the final decision of the classification. Finally, these strategies are evaluated by three data sets and the results show that DS theory can improve the accuracy obviously, and the strategies based on MSVM and DS theory is very fit for solving problems with small data sets.