|
Title
|
The performance quality of LR, SVM, and RF for earthquake-induced landslides susceptibility mapping incorporating remote sensing imagery
|
|
Type
|
JournalPaper
|
|
Keywords
|
Earthquake-inducedlandslide .Randomforest (RF) .Logistic regression(LR) .Supportvectormachine (SVM) .ROC curve
|
|
Abstract
|
An earthquake with Ms 7.0 (33.2° N, 103.8° E) occurred in Jiuzhaigou County of Sichuan Province in China on 8 August 2017. This earthquake triggered a large number of landslides in the study area. Although the susceptibility quality level index has improved, the high-quality assessments still have remained rare. We adopted three models, including the logistic regression (LR), support vector machine (SVM), and random forest (RF) to study the quality performance of the susceptibility distribution rule of earthquakes induced landslides. We used satellite images of before and after earthquakes and landslides as well. We used the area under receiver operating characteristic (ROC) curve (AUC) and ratio to evaluate the model’s accuracy and quality performance, including the mapping availability susceptibility assessment. This study reveals that RF has the highest ratio (2.07) as compared to the LR (1.78) and SVM (1.90). The result shows that RF has more potential to implement future experiments in Sichuan Province because of a better performance quality in the susceptibility assessment of landslides induced by earthquakes.
|
|
Researchers
|
Himan Shahabi (Not In First Six Researchers), Xin Yang (Fifth Researcher), Zili Lai (Fourth Researcher), Saeid Pirasteh (Third Researcher), Luyao Li (Second Researcher), Rui Liu (First Researcher)
|