Seismic Landslide Susceptibility Assessment Using Principal Component Analysis And Support Vector Machinepreprint
Аннотация: Abstract Seismic landslides are dangerous natural hazards, causing immense damage in terms of human lives and property. Susceptibility assessment of earthquake triggered landslide is the scientific premise and theoretical basis of disaster emergency management of engineering. The aim of this study is to applied the seismic landslide susceptibility model to Dayong Expressway in Chenghai area prone to frequent earthquakes. Support vector machine is used to establish the assessment model based on the data of 716 landslides caused by Ludian Ms6.5 earthquake in 2014. To improve the universality of the assessment model in different regions. Principal component analysis (PCA) is used for reducing the dimension of landslide conditioning factors and weaking difference of the regional characteristics between historical earthquake regions with Dayong expressway area. To applied the SVM model for seismic landslide susceptibility in Dayong Expressway region where the conditioning factors information is similar to Ludian area. Gutenberg-Richter model and Dieterich model are used to assume an earthquake in Chenghai area for landslide susceptibility assessment. Inverse distance weight (IDW) method is used for assessing the landslide risk class of Dayong Expressway. The results show that the “Very high” landslide susceptibility class account for 0.63% of Chenghai area. The seismic landslide has the most obvious impact on the middle 13 km section of Dayong expressway and this section account for 8.9% is defined as high-risk class. The study verifies the practicability of the seismic landslide susceptibility model based on machine learning and provides constructive reference for the susceptibility assessment of engineering facilities under earthquake.
Год издания: 2021
Авторы: Ziyao Xu, Ailan Che, Yanbo Cao, Fanghao Zhang
Издательство: Research Square (United States)
Источник: Research Square (Research Square)
Ключевые слова: Landslides and related hazards, Tree Root and Stability Studies, Yersinia bacterium, plague, ectoparasites research
Открытый доступ: green