Multi-class support vector machine for fault diagnosis
Multi-class support vector machine for fault diagnosis
复制标题
用于故障诊断的多类支持向量机
DOI：
--
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
Wang, Qiang
中科院分区：
文献类型：
--
作者：
Wang, Ting;Chen, Huan-Huan;Wang, Qiang
Decision tree support vector machine(SVM),which combine SVM and decision tree,is proposed for multi-class classification in order to perform the multi-class fault diagnosis tasks.Considering the limitations of the conventional methods,genetic algorithm,as a global randomized search method,was introduced into the formation of decision tree.By analyzing the distribution of the training samples,the fitness function of genetic algorithm was defined as the distance between the clustering centers of the two sub-classes so that the most separable classes could be separated at each node of decision tree.To testify the effectiveness of the proposed method,numerical simulations are conducted on three datasets compared with "one-against-one" and "one-against-all".The results show that the proposed method has better performance and higher generalization ability than the two conventional methods.
