“…Its main advantage is that it allows the use of a binary dependent variable-the occurrence of landslides in susceptibility mapping (Yilmaz, 2009;Ozdemir and Altural, 2013;Kavzoglu et al, 2014;Chan et al, 2018;Tian et al, 2019). In addition, methodological models, such as the frequency ratio model, multivariate adaptive regression splines, the generalized summation model, the deterministic factor method, the weight of evidence, and the entropy method have also been widely used in the spatial modeling of landslide hazards (Pardeshi et al, 2013;Xu et al, 2013;Jaafari et al, 2014;Regmi et al, 2014;Conoscenti et al, 2015;Youssef et al, 2015;Ilia and Tsangaratos, 2016;Xie et al, 2017;Ma and Xu, 2019). However, researchers have found that traditional mathematical methods were not sufficient to address the problems caused by the complexity of topography, geology, and other elements associated with the occurrence of landslide hazards: more flexible nonlinear methods were generally needed.…”