Risk assessment for debris flow by support vector machine

Debris flow is one of the most destructive natural hazards. The risk assessment of debris flow is important in disaster prevention. However, the relationship between debris flow and its determining factors is so complex and nonlinear that conventional mechanics methods can not be applied satisfactorily. Therefore, in this paper, the Support Vector Machine (SVM) method was employed to address this problem. SVM is a new creative learning system based oil the statistical learning theory. Using the SVM global optimization, solutions for problems with high dimension and nonlinearity call be found through small training samples. SVM can learn from case histories and then map a relationship between a debris flow and its determining factors. Once this relationship is built, it can be used to estimate the risk of other debris flows. Numerical results show that the SVM method is feasible and effective.