Evaluation of neural network models for landslide susceptibility assessment

被引:36
作者
Yi, Yaning [1 ,2 ,3 ]
Zhang, Wanchang [4 ]
Xu, Xiwei [1 ,2 ,3 ]
Zhang, Zhijie [5 ]
Wu, Xuan [4 ,6 ]
机构
[1] Minist Emergency Management China, Natl Inst Nat Hazards, Beijing, Peoples R China
[2] China Earthquake Adm, Key Lab Crustal Dynam, Beijing, Peoples R China
[3] Xichang Observ Nat Disaster Dynam Strike Slip Fau, Xichang, Peoples R China
[4] Chinese Acad Sci, Aerosp Informat Res Inst, Key Lab Digital Earth Sci, Beijing, Peoples R China
[5] Univ Connecticut, Dept Geog, Storrs, CT USA
[6] Univ Chinese Acad Sci, Beijing, Peoples R China
基金
中国国家自然科学基金;
关键词
Deep learning; landslide; data-driven; GIS; Sichuan; SAMPLING STRATEGIES; DECISION TREE; LAND-USE; GIS; CLASSIFICATION; ZONATION; HAZARD; REGION; FOREST; ISLAND;
D O I
10.1080/17538947.2022.2062467
中图分类号
P9 [自然地理学];
学科分类号
0705 ; 070501 ;
摘要
Identifying and assessing the disaster risk of landslide-prone regions is very critical for disaster prevention and mitigation. Owning to their special advantages, neural network algorithms have been widely used for landslide susceptibility mapping (LSM) in recent decades. In the present study, three advanced neural network models popularly used in relevant studies, i.e. artificial neural network (ANN), one dimensional convolutional neural network (1D CNN) and recurrent neural network (RNN), were evaluated and compared for LSM practice over the Qingchuan County, Sichuan province, China. Extensive experimental results demonstrated satisfactory performances of these three neural network models in accurately predicting susceptible regions. Specifically, ANN and 1D CNN models yielded quite consistent LSM results but slightly differed from those of RNN model spatially. Nevertheless, accuracy evaluations revealed that the RNN model outperformed the other two models both qualitatively and quantitatively but its complexity was relatively high. Experiments concerning training hyper-parameters on the performance of neural network models for LSM suggested that relatively small batch size values with Tanh activation function and SGD optimizer are essential to improve the performance of neural network models for LSM, which may provide a thread to help those who apply these advanced algorithms to improve their efficiency.
引用
收藏
页码:934 / 953
页数:20
相关论文
共 53 条
[1]   GIS-based landslide susceptibility modeling: A comparison between fuzzy multi-criteria and machine learning algorithms [J].
Ali, Sk Ajim ;
Parvin, Farhana ;
Vojtekova, Jana ;
Costache, Romulus ;
Nguyen Thi Thuy Linh ;
Quoc Bao Pham ;
Vojtek, Matej ;
Gigovic, Ljubomir ;
Ahmad, Ateeque ;
Ghorbani, Mohammad Ali .
GEOSCIENCE FRONTIERS, 2021, 12 (02) :857-876
[2]  
[Anonymous], 1996, TRANSP RES BOARD SPE
[3]   Landslide susceptibility estimation by random forests technique: sensitivity and scaling issues [J].
Catani, F. ;
Lagomarsino, D. ;
Segoni, S. ;
Tofani, V. .
NATURAL HAZARDS AND EARTH SYSTEM SCIENCES, 2013, 13 (11) :2815-2831
[4]   Performance evaluation of the GIS-based data mining techniques of best-first decision tree, random forest, and naive Bayes tree for landslide susceptibility modeling [J].
Chen, Wei ;
Zhang, Shuai ;
Li, Renwei ;
Shahabi, Himan .
SCIENCE OF THE TOTAL ENVIRONMENT, 2018, 644 :1006-1018
[5]   Landslide characteristics and, slope instability modeling using GIS, Lantau Island, Hong Kong [J].
Dai, FC ;
Lee, CF .
GEOMORPHOLOGY, 2002, 42 (3-4) :213-228
[6]   Landslide susceptibility zonation using analytical hierarchy process and GIS for the Bojnurd region, northeast of Iran [J].
Daneshvar, Mohammad Reza Mansouri .
LANDSLIDES, 2014, 11 (06) :1079-1091
[7]   Comparing the prediction performance of a Deep Learning Neural Network model with conventional machine learning models in landslide susceptibility assessment [J].
Dieu Tien Bui ;
Tsangaratos, Paraskevas ;
Viet-Tien Nguyen ;
Ngo Van Liem ;
Phan Trong Trinh .
CATENA, 2020, 188
[8]   A spatially explicit deep learning neural network model for the prediction of landslide susceptibility [J].
Dong Van Dao ;
Jaafari, Abolfazl ;
Bayat, Mahmoud ;
Mafi-Gholami, Davood ;
Qi, Chongchong ;
Moayedi, Hossein ;
Tran Van Phong ;
Hai-Bang Ly ;
Tien-Thinh Le ;
Phan Trong Trinh ;
Chinh Luu ;
Nguyen Kim Quoc ;
Bui Nhi Thanh ;
Binh Thai Pham .
CATENA, 2020, 188
[9]   Different sampling strategies for predicting landslide susceptibilities are deemed less consequential with deep learning [J].
Dou, Jie ;
Yunus, Ali P. ;
Merghadi, Abdelaziz ;
Shirzadi, Ataollah ;
Hoang Nguyen ;
Hussain, Yawar ;
Avtar, Ram ;
Chen, Yulong ;
Binh Thai Pham ;
Yamagishi, Hiromitsu .
SCIENCE OF THE TOTAL ENVIRONMENT, 2020, 720
[10]   An integrated artificial neural network model for the landslide susceptibility assessment of Osado Island, Japan [J].
Dou, Jie ;
Yamagishi, Hiromitsu ;
Pourghasemi, Hamid Reza ;
Yunus, Ali P. ;
Song, Xuan ;
Xu, Yueren ;
Zhu, Zhongfan .
NATURAL HAZARDS, 2015, 78 (03) :1749-1776