Entropy-based particle swarm optimization with clustering analysis on landslide susceptibility mapping

Generation of landslide susceptibility maps is important for engineering geologists and geomorphologists. The goal of this study is to generate a reliable susceptibility map based on digital elevation modeling and remote sensing data through clustering technique. This study focused on the landslide problems on a vast area located at Shei Pa National Park, Miao Li, Taiwan. Two stages of analysis were used to extract the dominant attributes and thresholds: (1) calculate the entropy with regard to the measure of influenced variables to the occurrence of landslide and (2) use the clustering analysis K-means with particle swarm optimization (KPSO) to resolve the difficulties in creating landslide susceptibility maps. The knowledge scope with regard to core factors and thresholds are solved. The self-organization map (SOM) is used as a parallel study for comparison. The overall accuracy of the susceptibility map is 86 and 77 % for KPSO and SOM, respectively. Then, the susceptibility maps are drawn and verifications made. The generation of a susceptibility map is useful for decision makers and managers to handle the landslide risk area.