Main content area

A novel hybrid KPCA and SVM with PSO model for identifying debris flow hazard degree: a case study in Southwest China

Qian, Xin, Chen, Jian-Ping, Xiang, Liang-Jun, Zhang, Wen, Niu, Cen-cen
Environmental earth sciences 2016 v.75 no.11 pp. 991
case studies, models, principal component analysis, support vector machines, China
A novel support vector machine (SVM) model that combines kernel principal component analysis (KPCA) with particle swarm optimization (PSO) is proposed for identifying debris flow hazard degree. The proposed model consists of three stages. First, KPCA is used to extract nonlinear feature information and solve the linear correlation of input data. Second, PSO is implemented to optimize the selection of the penalty factor and the kernel function parameter of the SVM classifier in the solution space. This optimization can prevent blindness in parameter selection and improve the identification accuracy of the model. Finally, the extracted new features and the optimal parameters for the SVM classifier are employed to predict the test data. Experimental results show that the proposed model can more effectively select discriminating input features and achieve higher identification accuracy than other hybrid models. Therefore, the proposed model exhibits high potential to become a useful tool for identifying debris flow hazard degree in limited samples.