Secondary-factor division optimization based on ET-GA to improve the accuracy of LSA obtained using SSA-DBN compared to the original division method: a case study in southern Sichuan, China

Landslide susceptibility assessment (LSA) is an essential method to prevent and mitigate landslides. Many studies thus involve the LSA improve accuracy of using various methods; however, most optimize by improving evaluation models neglect influence factors. The present study initially selects fifteen factors combines optimization algorithms, namely, genetic algorithm (GA), with entropy theory (ET) divide division results are compared that natural breakpoint (NBM) equidistant empirical (EEM) which commonly used methods. employed select based on chi-square test multicollinearity test, optimal number involved in factor systems determined recursive feature elimination out-of-bag error. LSAs ET-GA, NBM, EEM then obtained improved deep belief networks (DBN), distribution patterns landslide maps similar. density very high three methods more than times low susceptibility, illustrates applicable area. evaluated area under receiver operating characteristic curve (AUC). ET-GA showed highest AUC, proving algorithms can for