Landslide risk evaluation in Shenzhen based on stacking ensemble learning and InSAR

Construction activities of accelerated urbanization in Shenzhen have increased the landslide risk area, which has intensified potential threat to human and natural environment. However, landslides is poorly evaluated. In this paper, a evaluation (LRE) model constructed using susceptibility map (LSM) vulnerability. experiment, stacking ensemble learning (SEL) based on convolutional neural network (CNN), multilayer perceptron (MLP), gated recurrent unit (GRU) support vector machine regression (SVR) generate LSM by topography, geology, engineering activities, time-series precipitation normalized difference vegetation index (NDVI). Road network, building distribution density annual average data are used evaluate vulnerability entropy weight method. study, multiple statistical indicators performance model, Interferometric Synthetic Aperture Radar (InSAR) deformation utilized verify LRE results Shenzhen. The show that SEL method more refined for LSM, with best overall accuracy, especially receiver operating characteristic curve (ROC), where accuracy improved nearly 8%. Shenzhen, very high, moderate, low areas account 0.283%, 0.451%, 0.859%, 36.890% 61.517%, respectively. most high InSAR clear concentrated trend large rate. Research can provide technical disaster prevention