Landslide Susceptibility Evaluation Using Multiple Machine Learning Models: A Case Study of Wanzhou District in the Three Gorges Reservoir Area

ABSTRACT Wanzhou District, located in the heart of the Three Gorges Reservoir Area, is highly prone to landslide disasters due to its complex geological conditions, intense rainfall, and fluctuations in reservoir water levels. To aid effective disaster prevention and mitigation efforts, this study assesses landslide susceptibility based on 715 landslide events collected from recent field investigations and statistical surveys. Principal component analysis (PCA) and Pearson's correlation coefficient (PCC) methods were employed to optimise the factor system, ultimately selecting 12 key influencing factors. Elevation was excluded due to its high correlation with annual rainfall (| r | = 0.88) and its relatively lower weight (0.0785, compared with 0.1133 for rainfall). Statistical analysis indicated that landslides are primarily concentrated within 100–300 m of rivers and 400–500 m of residential buildings, with 62.4% of landslides closely associated with anthropogenic engineering activities. Three machine learning models—random forest (RF), multilayer perceptron combined with Bagging (MLP‐Bagging) and extreme gradient boosting (XGBoost)—were applied for susceptibility modeling. The RF model achieved the highest area under the curve (AUC), with a value of 0.887 and identified high and very high susceptibility zones that accounted for 72.87% of historical landslides, outperforming the other models. Zoning results based on the natural breaks method revealed extremely high susceptibility zones along the Yangtze River and urban areas, reflecting the combined impacts of fluctuating reservoir water levels and urbanisation. Overall, the RF model demonstrated superior performance in landslide susceptibility assessment and the incorporation of up‐to‐date landslide records significantly enhanced the reliability and practical value of the results, providing scientific support for disaster prevention and mitigation in the Three Gorges Reservoir Area.