Decision Tree Twin Support Vector Machine Based on Kernel Clustering for Multi-class Classification
To deal with multi-class classification problems using twin support vector machines (TWSVMs), this paper proposes a novel multi-class classifier, decision tree twin support vector machine based on kernel clustering (DT2SVM-KC). We employ the kernel clustering algorithm to generate a binary tree, and for each non-leaf node, we obtain a pair of non-parallel hyperplanes by using TWSVM. Simulation results show that the proposed method can keep the strength of decision tree in computation time and has better performance on most used datasets compared with other multi-class classification methods based on TWSVMs.
KeywordsTwin support vector machine Decision tree Multi-class classification Kernel function
We thanks to all those who contribute to the data sets of UCI machine learning repository that are used in this paper. This work was supported in part by the National Natural Science Foundation of China under Grant Nos. 61373093, 61402310, 61672364 and 61672365, by the Soochow Scholar Project of Soochow University, by the Six Talent Peak Project of Jiangsu Province of China, and by the Graduate Innovation and Practice Program of colleges and universities in Jiangsu Province.
- 6.Sartakhti, J.S., Ghadiri, N., Afrabandpey, H., et al: Fuzzy least squares twin support vector machines. Artificial Intelligence (2015)Google Scholar
- 13.Zhen, W., Jin, C., Ming, Q.: Non-parallel planes support vector machine for multi-class classification. In: International Conference on Logistics Systems and Intelligent Management, pp. 581–585. IEEE, Harbin (2010)Google Scholar