Abstract :
Debris flow hazard assessment is a basic work of hazard monitoring, forecast, alleviation and control. Seven factors, including the maximum volume of once flow (LI), occurrence frequency of debris flow (L2), watershed area (SI), main channel length (S2), watershed relative height difference (S3), valley incision density (S6) and the length ratio of sediment supplement (S9) are chosen as evaluation factors of debris flow hazard degree. Using support vector machine (SVM) theory, 259 basic data of 37 debris flow channels in Yunnan Province are selected as learning samples in this study , then a kind of debris flow hazard assessment model based on SVM is produced. First instance applications gave encouraging results. After Cross Validation test, accuracy of this model came to 70.00%. Through verifying 7 groups of test data, classification accuracy came to 85.71%. The model shows that it has the advantages of best generation, convenience and high precision. SVM is regarded as a broadly applicative tool in debris flow hazard assessment.
Keywords :
disasters; environmental factors; erosion; floods; geophysics computing; hydrological techniques; support vector machines; SVM; Yunnan province; cross validation test; debris flow channels; debris flow hazard assessment; debris flow occurrence frequency; learning samples; main channel length; model accuracy; once flow maximum volume; sediment supplement length ratio; support vector machine; valley incision density; watershed area; watershed relative height difference; Condition monitoring; Economic forecasting; Educational institutions; Hazards; Input variables; Machine learning; Support vector machine classification; Support vector machines; Telecommunication computing; Testing;
