Landslide Susceptibility Assessment Based on Different MaChine Learning Methods in Zhaoping County of Eastern Guangxi

Regarding the ever increasing and frequent occurrence of serious landslide disaster in eastern Guangxi, current study was implemented to adopt support vector machines (SVM), particle swarm optimization (PSO-SVM), random forest (RF), (PSO-RF) methods assess susceptibility Zhaoping County. To this end, 10 disaster-related variables including digital elevation model (DEM)-derived, meteorology-derived, Landsat8-derived, geology-derived, human activities factors were provided. Of 345 locations found, 70% used train models, rest them performed for verification. The aforementioned four models run, evaluation maps produced. Then, receiver operating characteristics (ROC) curves, statistical analysis, field investigation test verify efficiency these models. Analysis comparison results denoted that all well as indicated by area under curve (AUC) values ROC curves from 0.863 0.934. Among them, it has been shown PSO-RF highest accuracy other followed PSO-SVM model, RF SVM model. Moreover, also showed PSO algorithm a good effect on Furthermore, devolved present are promising could be transferred regions evaluation. In addition, can provide suggestions reduction prevention County Guangxi.