基于机器学习的滑坡易发性预测建模及其主控因子识别

被引：38

作者：

黄发明 ^{[1
]}

胡松雁 ^{[1
]}

闫学涯 ^{[1
]}

李明 ^{[1
]}

王俊宇 ^{[1
]}

李文彬 ^{[1
]}

郭子正 ^{[2
]}

范文彦 ^{[1
]}

机构：

[1] 南昌大学建筑工程学院

[2] 中国地质大学(武汉)工程学院

来源：

地质科技通报 | 2022年 / 41卷 / 02期

基金：

国家重点研发计划;

关键词：

滑坡易发性预测; 不确定性分析; 主控因子识别; 支持向量机; 随机森林;

D O I：

10.19509/j.cnki.dzkq.2021.0087

中图分类号：

P642.22 [滑坡];

学科分类号：

0837 ;

摘要：

不同机器学习预测滑坡易发性的建模过程及其不确定性有所差异，另外如何有效识别滑坡易发性的主控因子意义重大。针对上述问题，以支持向量机(support vector machine,简称SVM)和随机森林(random forest,简称RF)为例探讨了基于机器学习的滑坡易发性预测及其不确定性，创新地提出了“权重均值法”来综合计算出更准确的滑坡主控因子。首先获取陕西省延长县滑坡编录和10类基础环境因子，将因子频率比值作为SVM和RF的输入变量；再将滑坡与随机选择的非滑坡样本划分为训练集和测试集，用训练好的机器学习预测出滑坡易发性并制图；最后用受试者工作曲线、均值和标准差等来评估建模不确定性，并计算滑坡主控因子。结果表明：(1)机器学习能有效预测出区域滑坡易发性，RF预测的滑坡易发性精度高于SVM,而其不确定性低于SVM,但两者的易发性分布规律整体相似；(2)权重均值法计算出延长县滑坡主控因子依次是坡度、高程和岩性。实例分析和文献综述显示RF模型相较于其他机器学习模型属于可靠性较高的易发性模型。

引用

页码：79 / 90

页数：12

共 36 条

[11] Coseismic landslides induced by the 2018 M w 6.6 Iburi; Japan; Earthquake: spatial distribution; key factors weight; and susceptibility regionalization.[J].Ming Chang;Yu Zhou;Chao Zhou;Tristram C. Hales.Landslides.2020, 2
[12] Landslide susceptibility prediction based on a semi-supervised multiple-layer perceptron model
Huang, Faming
Cao, Zhongshan
Jiang, Shui-Hua
Zhou, Chuangbing
Huang, Jinsong
Guo, Zizheng
[J]. LANDSLIDES, 2020, 17 (12) : 2919 - 2930
[13] Landslide Susceptibility Prediction Considering Regional Soil Erosion Based on Machine-Learning Models
Huang, Faming
Chen, Jiawu
Du, Zhen
Yao, Chi
Huang, Jinsong
Jiang, Qinghui
Chang, Zhilu
Li, Shu
[J]. ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION, 2020, 9 (06)
[14] Comparing the prediction performance of a Deep Learning Neural Network model with conventional machine learning models in landslide susceptibility assessment.[J].Dieu Tien Bui;Paraskevas Tsangaratos;Viet-Tien Nguyen;Ngo Van Liem;Phan Trong Trinh.Catena.2020,
[15] Landslide Susceptibility Prediction Modeling Based on Remote Sensing and a Novel Deep Learning Algorithm of a Cascade-Parallel Recurrent Neural Network
Zhu, Li
Huang, Lianghao
Fan, Linyu
Huang, Jinsong
Huang, Faming
Chen, Jiawu
Zhang, Zihe
Wang, Yuhao
[J]. SENSORS, 2020, 20 (06)
[16] Landslide Susceptibility Prediction Based on Remote Sensing Images and GIS: Comparisons of Supervised and Unsupervised Machine Learning Models
Chang, Zhilu
Du, Zhen
Zhang, Fan
Huang, Faming
Chen, Jiawu
Li, Wenbin
Guo, Zizheng
[J]. REMOTE SENSING, 2020, 12 (03)
[17] Machine learning methods for landslide susceptibility studies: A comparative overview of algorithm performance.[J].Abdelaziz Merghadi;Ali P. Yunus;Jie Dou;Jim Whiteley;Binh ThaiPham;Dieu Tien Bui;Ram Avtar;Boumezbeur Abderrahmane.Earth-Science Reviews.2020, prepublish
[18] Prediction of soil water retention curve using Bayesian updating from limited measurement data.[J].Weiping Liu;Xiaoyan Luo;Faming Huang;Mingfu Fu.Applied Mathematical Modelling.2019,
[19] Coupling logistic model tree and random subspace to predict the landslide susceptibility areas with considering the uncertainty of environmental features
Luo, Xiangang
Lin, Feikai
Chen, Yihong
Zhu, Shuang
Xu, Zhanya
Huo, Zhibin
Yu, Mengliang
Peng, Jing
[J]. SCIENTIFIC REPORTS, 2019, 9 (1)
[20] Landslide Susceptibility Prediction Using Particle-Swarm-Optimized Multilayer Perceptron: Comparisons with Multilayer-Perceptron-Only, BP Neural Network, and Information Value Models
Li, Deying
Huang, Faming
Yan, Liangxuan
Cao, Zhongshan
Chen, Jiawu
Ye, Zhou
[J]. APPLIED SCIENCES-BASEL, 2019, 9 (18):

← 1 2 3 4 →