The survival analysis of landslide dams based on database imputation

Catastrophic floods resulting from landslide dam breaching have profound impacts on the lives of residents, infrastructure, and ecological environment in downstream regions. Investigating survival mechanisms at various stages following landslide dam formation is advantageous for disaster prevention and mitigation efforts. In this study, different types of factors are firstly imputed using three methods, i.e., linear regression (LR), K-nearest neighbors (KNN), and random forest (RF). The Kaplan–Meier (K-M) method is then employed to assess the long-term survival of the landslide dam. The using factors are classified via K-means. Finally, the second landslide dam of Baige is introduced to validate the reliability of K-M curve. The results indicate that the clustering method significantly enhances the differentiation between various grades. The survival of the landslide dam in the early stages after formation is significantly impacted by hydrological parameters and geological and climatic conditions, particularly geomorphic characteristics. The effects of these factors are observed to gradually weaken over time.