Information-based dichotomization: A method for multiclass Support Vector Machines

Songsiri, Patoomsiri; Kijsirikul, Boonserm; Phetkaew, Thimaporn

doi:10.1109/ijcnn.2008.4634264

Cited by 4 publications

(3 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As described in Algorithm 2, IB-DTree constructs a tree using a recursive procedure starting from the root node from lines 1-7 with all candidate classes in line 2. The node-adding procedure will be processed from lines [8][9][10][11][12][13][14][15][16][17][18][19]. First, the initial classifier h with the lowest entropy will be selected.…”

Section: The Information-based Decision Treementioning

confidence: 99%

“…The Information-Based Dichotomization (IBD) [18], proposed by Songsiri et al, employs information theory to construct a multi-class classification tree. In each node, IBD selects the OVO classifier with minimum entropy.…”

Section: Information-based Dichotomizationmentioning

confidence: 99%

“…Fei and Liu have proposed the Binary Tree of SVM [9] that selects tree node classifiers randomly or selects by the training data centroids. Songsiri et al have proposed the Information-Based Dichotomization Tree [18] that uses entropy for classifier selection. Chen et al employ the Adaptive Binary Tree [6] that applies the minimization of the average number of support vectors for tree construction.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Efficient Decision Trees for Multi–Class Support Vector Machines Using Entropy and Generalization Error Estimation

Kantavat

Kijsirikul

Songsiri

et al. 2018

International Journal of Applied Mathematics and Computer Science

Self Cite

View full text Add to dashboard Cite

We propose new methods for Support Vector Machines (SVMs) using tree architecture for multi-class classification. In each node of the tree, we select an appropriate binary classifier using entropy and generalization error estimation, then group the examples into positive and negative classes based on the selected classifier and train a new classifier for use in the classification phase. The proposed methods can work in time complexity between O(log 2 N ) to O(N ) where N is the number of classes. We compared the performance of our proposed methods to the traditional techniques on the UCI machine learning repository using 10-fold cross-validation. The experimental results show that our proposed methods are very useful for the problems that need fast classification time or problems with a large number of classes as the proposed methods run much faster than the traditional techniques but still provide comparable accuracy.

show abstract