Landslide susceptibility prediction using C5.0 decision tree model

Regional landslide susceptibility prediction (LSP) research is of great significance to the prevention and control landslides. This study focuses on LSP modelling based decision tree model. Taking northern part An’yuan County Jiangxi Province as an example, 14 environmental factors including elevation, gully density lithology are obtained geographical information system (GIS) remote sensing satellite. Frequency Ratio method C5.0 (DT) model coupled build DT for modelling. Then predicted results graded into five attribute intervals. Finally, performance evaluated by comparing area value under receiver operating characteristic curve (ROC) classification susceptibility. The show that AUC accuracy 0.805, consistent with actual distribution pattern landslides in this County.