Regional Landslide Susceptibility Mapping Based on Grey Relational Degree Model

Statistical and machine learning models, such as support vector machine (SVM), have been widely used to assess the landslide susceptibility. However, the modeling processes of statistical and machine learning model are generally complex.For example, it is difficult to select reasonable non-landslide grid cells when the machine learning models are trained and tested, and many model parameters need to be determined.In order to improve the efficiency and accuracy of the model used for landslide susceptibility assessment, the grey relational degree (GRD) model is proposed. The GRD model can efficiently calculate the quantitative relational degrees between the comparative samples and the reference sample, and it has the advantages of simple modeling process and accurate assessment results.However, few studies have been done on the GRD model.In this study, the GRD model is used to assess the landslide susceptibility in the Nantian and Yamei maps (Nantian area) in the Feiyunjiang River basin, Zhejiang Province of China, and the assessment results of the GRD model are compared with those of the SVM model. The results show that the GRD model has higher prediction rate than the SVM model in the high and very high susceptibility areas, and has slightly lower prediction rate than the SVM in the moderate susceptibility area. On the whole, the GRD model has slightly higher prediction rate than the SVM for landslide susceptibility assessment in Nantian area. Meanwhile, the results also show that the model process of GRD is simple, it has higher efficiency than the SVM. The GRD model provides a novel idea for landslide susceptibility assessment.