Spatial landslide susceptibility assessment using machine learning techniques assisted by additional data created with generative adversarial networks

作     者：Husam A.H.Al-Najjar Biswajeet Pradhan Husam A.H. Al-Najjar;Biswajeet Pradhan

作者机构：Centre for Advanced Modelling and Geospatial Information Systems(CAMGIS)Faculty of Engineering and ITUniversity of Technology Sydney2007NSWAustralia Department of Energy and Mineral Resources EngineeringSejong UniversityChoongmu-gwan209Neungdong-roGwangin-guSeoul05006Republic of Korea Center of Excellence for Climate Change ResearchKing Abdulaziz UniversityP.O.Box 80234Jeddah21589Saudi Arabia Earth Observation CenterInstitute of Climate ChangeUniversiti Kebangsaan Malaysia43600UKMBangiSelangorMalaysia 

出 版 物：《Geoscience Frontiers》 (地学前缘（英文版）)

年 卷 期：2021年第12卷第2期

页      面：625-637页

核心收录：

学科分类：081803[工学-地质工程] 08[工学] 0818[工学-地质资源与地质工程] 

基　　金：This research is funded by the Centre for Advanced Modeling and Geospatial Information Systems(CAMGIS) Faculty of Engineering and Information Technology the University of Technology Sydney Australia 

主　　题：Landslide susceptibility Inventory Machine learning Generative adversarial network Convolutional neural network Geographic information system 

摘      要：In recent years,landslide susceptibility mapping has substantially improved with advances in machine ***,there are still challenges remain in landslide mapping due to the availability of limited inventory *** this paper,a novel method that improves the performance of machine learning techniques is *** proposed method creates synthetic inventory data using Generative Adversarial Networks(GANs)for improving the prediction of *** this research,landslide inventory data of 156 landslide locations were identified in Cameron Highlands,Malaysia,taken from previous projects the authors worked ***,slope,aspect,plan curvature,profile curvature,total curvature,lithology,land use and land cover(LULC),distance to the road,distance to the river,stream power index(SPI),sediment transport index(STI),terrain roughness index(TRI),topographic wetness index(TWI)and vegetation density are geo-environmental factors considered in this study based on suggestions from previous works on Cameron *** show the capability of GANs in improving landslide prediction models,this study tests the proposed GAN model with benchmark models namely Artificial Neural Network(ANN),Support Vector Machine(SVM),Decision Trees(DT),Random Forest(RF)and Bagging ensemble models with ANN and SVM *** models were validated using the area under the receiver operating characteristic curve(AUROC).The DT,RF,SVM,ANN and Bagging ensemble could achieve the AUROC values of(0.90,0.94,0.86,0.69 and 0.82)for the training;and the AUROC of(0.76,0.81,0.85,0.72 and 0.75)for the test,*** using additional samples,the same models achieved the AUROC values of(0.92,0.94,0.88,0.75 and 0.84)for the training and(0.78,0.82,0.82,0.78 and 0.80)for the test,*** the additional samples improved the test accuracy of all the models except *** a result,in data-scarce environments,this research showed that utilizing GANs to generate supplementary samples is