Susceptibility Analysis of Earthquake-Induced Landslide Using Random Forest Method

Abstract

This paper used random forest method to produce susceptibility analysis for earthquake-induced landslides. A landslide susceptibility model was built based on random forest method and data of environment variables and Newmark displacement. Landslide susceptibility mapping and attributes evaluation was also preformed. This model was tested in Pingwu County of Sichuan Province. The results show that the random forest method provides better accuracy on landslide prediction, and also shows smoother characteristic on the distribution of the landslide susceptibility when compared to wildly used regression models. The results confirm that the random forest method can be performed as an effective approach for landslide susceptibility evaluation and mapping.