Can different machine learning methods have consistent interpretations of DEM-based factors in shallow landslide susceptibility assessments?

Research on the application of machine learning (ML) models to landslide susceptibility assessments has increased in popularity in recent years and has focused mostly on topographic factors from digital elevation models (DEMs). However, few studies have focused on the explanatory effects of these factors on different models, i.e., whether DEM-based factors affect different models in the same way. This study investigated whether different ML models could have consistent interpretations of DEM-based factors with the help of explainable algorithms. Six ML models, including a support vector machine (SVM), a neural network (NN), extreme gradient boosting (XGB), a random forest (RF), linear regression (LR), and K-nearest neighbors (KNN), were trained and evaluated on five geospatial datasets derived from different DEMs. Each dataset contained 14 conditional factors (8 DEM-based and 6 non-DEM-based) of 8,912 landslide samples. Model performance was evaluated by using the accuracy, precision, recall rate, F1-score, and the kappa coefficient, as well as receiver operating characteristic curves. Furthermore, explainable analyses including Shapley additive explanations and partial dependence plots, were employed to investigate the effects of topographic factors on landslide susceptibility. This study can benefit the broader landslide research community by clarifying how topographic factors affect ML models for facilitating disaster prevention and control.