Landslide Susceptibility Prediction Based on High-Trust Non-Landslide Point Selection

Landslide susceptibility prediction has the disadvantages of being challenging to apply expanding landslide samples and low accuracy a subjective random selection non-landslide samples. Taking Fu’an City, Fujian Province, as an example, model based on semi-supervised framework using particle swarm optimization optimize extreme learning machines (SS-PSO-ELM) is proposed. Based samples, constructed through Density Peak Clustering (DPC), Frequency Ratio (FR), Random Forest (RF) models expand divide sample data. The was predicted high-trust data input variables data-driven model. results show that area under curve (AUC) valued at SS-PSO-ELM for 0.893 root means square error (RMSE) 0.370, which better than ELM PSO-ELM without framework. It shows more effective in susceptibility. Thus, it provides new research idea predicting