Slope Unit-Based Landslide Susceptibility Mapping Using Certainty Factor, Support Vector Machine, Random Forest, CF-SVM and CF-RF Models

Landslide susceptibility mapping is very important for landslide risk evaluation and land use planning. Toward this end, paper presents a case study in Ningqiang County, Shanxi Province, China. Slope units were selected as the basic units. A traditional statistical certainty factor model (CF), machine learning support vector (SVM) random forest (RF), along with hybrid CF-SVM CF-RF applied to analyze susceptibility. Firstly, 10 conditioning factors selected, namely slope-angle, altitude, slope aspect, degree of relief, lithology, distance rivers, faults, roads, average annual rainfall normalized difference vegetation index. The 23,169 generated from Digital Elevation Model corresponding layers produced both geological geographical data. Then, was carried out using five models, respectively. Next, density (LD), frequency ratio (FR), area under curve (AUC) other indicators used validate rationality, performance accuracy models. results showed that maps different models all reasonable. In each map, LD FR greatest zones classed having high susceptibility, followed by high, moderate, low classes, From comparison ROC curves, RF based on most appropriate area. It also found combination weaker learner (CF here) stronger (SVM can impact applicability model.