Support vector machine tree based on feature selection

Authors:
Qinzhen Xu;Wenjiang Pei;Luxi Yang;Zhenya He
Affiliations:
Dept. Radio Engineering, Southeast University, Nanjing, China;Dept. Radio Engineering, Southeast University, Nanjing, China;Dept. Radio Engineering, Southeast University, Nanjing, China;Dept. Radio Engineering, Southeast University, Nanjing, China
Venue:
ICONIP'06 Proceedings of the 13 international conference on Neural Information Processing - Volume Part I
Year:
2006

Citing 8
Cited 2

Fast training of support vector machines using sequential minimal optimization

Advances in kernel methods
Enlarging the Margins in Perceptron Decision Trees

Machine Learning
Asymptotic behaviors of support vector machines with Gaussian kernel

Neural Computation
Improving learning accuracy of fuzzy decision trees by hybrid neural networks

IEEE Transactions on Fuzzy Systems
Decision trees can initialize radial-basis function networks

IEEE Transactions on Neural Networks
ANN-DT: an algorithm for extraction of decision trees from artificial neural networks

IEEE Transactions on Neural Networks
A comparison of methods for multiclass support vector machines

IEEE Transactions on Neural Networks
Feature selection in MLPs and SVMs based on maximum output information

IEEE Transactions on Neural Networks

A Facial Expression Recognition Approach Based on Novel Support Vector Machine Tree

ISNN '07 Proceedings of the 4th international symposium on Neural Networks: Advances in Neural Networks, Part III
Credit risk evaluation with kernel-based affine subspace nearest points learning method

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

A hybrid learning approach named confusion-cross-based support vector machine tree (CSVMT) has been proposed in our current work. It is developed to achieve a better performance for complex distribution problems even when the two parameters of SVM are not appropriately selected. One problem remained is that the trained internal nodes may be high complex for those high-dimensional feature space problems due to undesirable complexity added to the underlying probability distribution of the concept label for learning algorithm to capture –thus, learning models with high complexity are likely to depress the test efficiency and performance. In this paper, we proposed a feature selection based CSVMT (FS-SCVMT) learning approach in which the input space for each internal node is adaptively dimensionality reduced by sensitivity based feature selection. Experimental results showed that FS-SCVMT approach performed well.