State Classification Algorithm for Bus Based on Hierarchical Support Vector Machine

In this paper, with "One-against-rest" and "Oneagainst-one" as the foundation, the multi-classification algorithm is established on hierarchical support vector machines for the buses state classification. The multi-class classifier model is proposed on binary tree SVM to analyze the logic of the bus state. Then it used the date from Zhongshan, Guangdong intelligent traffic GPS to evaluate the classification accuracy level of the new multi-classification algorithm. The results show that the hierarchy SVM of buses multiclass has good results.