Full-Text Search:
Home|About CNKI|User Service|中文
Add to Favorite Get Latest Update

Collapse susceptibility mapping based on fuzzy clustering SVM:Take Anze County as an example

CAI Xinwei;No.149 Team of Gansu Coalfield Geology Bureau;  
In the collapse susceptibility mapping model, the random selection of non-disaster points will reduce the sample quality and affect the model training accuracy. This paper proposes a fuzzy clustering based support vector machine(SVM) evaluation method. First, the fuzzy clustering is used in the selection of non-disaster points for training samples to maximize the inter-class difference between disaster points and non-disaster points. Then the division coefficient is used to judge the degree of separation between the classification clusters and select non-disaster points samples under the optimal division coefficient. Finally, the disaster susceptibility evaluation is carried out using SVM. The results show that in the fuzzy clustering SVM integrated model, the point density of the disaster points in the extremely high-risk area increases by 0.381 7/km~2 compared with the single SVM model. The increase is significant, the regional distribution is more intensive and it is more consistent with the actual situation. The Receiver Operator Characteristic(ROC) curve is used to verify the results, and the Area Under Curve(AUC) is increased from 0.785 to 0.957. This shows that the fuzzy clustering SVM model has higher prediction accuracy for the evaluation of collapse susceptibility, which can provide a basis for the collapse risk assessment and management.
Download(CAJ format) Download(PDF format)
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.
©CNKI All Rights Reserved