Uncertain GM-CFSFDP Clustering Algorithm for Landslide Hazard Prediction

Uncertain GM-CFSFDP Clustering Algorithm for Landslide Hazard Prediction

Due to difficulties in obtaining and effectively processing rainfall in landslide hazard prediction, as well as the existing limitation in dealing with large-scale data sets in clustering by Fast Search and Find of Density Peaks (CFSFDP) algorithm, a novel CFSFDP algorithm based on grid and merging clusters (GM-CFSFDP) has been proposed to assess landslide susceptibility model. Firstly, this method adopted a new two-phase clustering algorithm, which is suitable for large-scale data sets. Secondly, the uncertain data model is presented to effectively quantify triggering factors (precipitation). At the same time, a novel Euclidean distance formula based on midpoint and length of uncertain data (E−ML distance formula) is designed, which makes the new method to manage the uncertain data. Finally, the prediction model of landslide hazards was constructed and verified in Baota district of Yan’an city. The experimental results show that the uncertain GM-CFSFDP clustering algorithm can effectively improve the accuracy of landslide hazard prediction.