A Comparative Study of Landslide Susceptibility Evaluation Based on Three Different Machine Learning Algorithms
-
摘要:
准确的滑坡易发性评价结果是山区滑坡灾害防治的关键,可有效规避潜在滑坡带来的风险。为获得准确、可靠的滑坡预防参考,笔者以云南芒市为研究对象,选取高程、地层岩性、年均降雨量等9项评价因子,通过多重共线性分析,构建研究区滑坡易发性评价指标体系。分别基于支持向量机(SVM)、BP神经网络和随机森林(RF)3种典型机器学习算法进行滑坡易发性评价。利用准确性(ACC)、ROC曲线下面积(AUC)、滑坡比(Sei)及野外实地考察对模型评价结果精度进行对比验证分析。结果显示RF模型的ACC、AUC和极高易发区的SeV值最高,分别为0.867、0.94、9.21;BP神经网络模型次之,其SeV值分别为0.829、0.90、9.14;SVM最低,其SeV值分别为0.794、0.88、6.85。此外,RF算法所得结果还与实地考察情况保持了较高的一致性。实验结果表明与其他两种算法相比,RF算法在芒市区域具有更高的准确性和可靠性,更适合用于该区域的滑坡易发性建模,且利用该模型获得的评价结果,能够为芒市区域的滑坡防治提供理论依据和科学参考。
Abstract:Accurate landslide susceptibility evaluation results are the key to landslide disaster prevention and control in mountainous areas, which can effectively avoid the risk caused by potential landslides. To obtain an accurate and reliable reference for landslide prevention, this paper selects nine evaluation factors, including elevation, stratigraphic lithology, average annual rainfall et al, and constructs a landslide susceptibility evaluation index system in the study area through multiple covariance analysis, taking Mangcheng, Yunnan Province as the research object. Subsequently, three typical machine learning models based on support vector machine (SVM), BP neural network and random forest (RF) were used for landslide susceptibility evaluation. Finally, the accuracy of the model evaluation results was compared and validated by using accuracy (ACC), area under the ROC curve (AUC), landslide ratio (Sei) and field fieldwork. The results showed that the RF model had the highest SeV values of 0.867, 0.94 and 9.21 for ACC, AUC, and very high susceptibility areas, respectively; the BP neural network model had the second highest values of 0.829, 0.90 and 9.14; the SVM had the lowest values of 0.794, 0.88 and 6.85; and the RF model results were more consistent with the field study. The results of experiments show that compared with the other two algorithms, the RF algorithm has higher accuracy and reliability in the Mangshi region and is more suitable for landslide susceptibility modeling in the region, and the evaluation results obtained by using the model can provide a theoretical basis and scientific reference for landslide control in the Mangshi region.
-
Key words:
- SVM /
- BP neural networks /
- RF /
- landslide susceptibilit /
- Mangshi city
-
-
表 1 评价因子多重共线性分析结果
Table 1. Results of multiple covariance analysis of evaluation factors
评价因子 容差 VIF 评价因子 容差 VIF 高程 0.781 1.281 起伏度 0.176 5.693 坡度 0.159 6.298 地层岩性 0.990 1.010 坡向 0.979 1.022 年均降雨量 0.981 1.019 平面曲率 0.708 1.413 土地利用 0.984 1.017 剖面曲率 0.869 1.151 表 2 测试样本精度评价
Table 2. Test sample accuracy evaluation
评价指标 评价模型 SVM BP神经网络 RF TP(真阳性) 130 143 138 FP(假阳性) 29 31 29 TN(真阴性) 139 138 156 FN(假阴性) 41 27 16 ACC(准确度) 0.794 0.829 0.867 表 3 易发性分区合理性检验结果
Table 3. Rationality test results of susceptibility zoning
评价
模型易发区 分级面
积/km2滑坡
点/个Mei Dei Sei SVM 极低(I) 1260.04 47 43.51% 8.32% 0.19 低(II) 624.7 47 21.57% 8.32% 0.39 中(III) 449.2 49 15.51% 8.67% 0.56 高(IV) 328.04 109 11.33% 19.29% 1.70 极高(V) 234.11 313 8.08% 55.40% 6.85 BP神经
网络极低(I) 1282.66 27 44.29% 4.78% 0.11 低(II) 619.45 31 21.39% 5.49% 0.26 中(III) 443.81 42 15.32% 7.43% 0.49 高(IV) 330.44 73 11.41% 12.92% 1.13 极高(V) 219.73 392 7.59% 69.38% 9.14 RF 极低(I) 1262.13 17 43.58% 3.00% 0.07 低(II) 684.86 20 23.65% 3.54% 0.15 中(III) 422.73 37 14.60% 6.55% 0.45 高(IV) 287.12 61 9.91% 10.80% 1.09 极高(V) 239.25 430 8.26% 76.11% 9.21 -
[1] 陈玉萍, 袁志强, 周博, 等. 遗传算法优化BP网络在滑坡灾害预测中的应用研究[J]. 水文地质工程地质, 2012, 39(1): 114-119
Chen Yuping, Yuan Zhiqiang, Zhou Bo, et al. Application of back propagation neural networks with optimization of genetic algorithms to landslide hazard prediction[J]. Hydrogeology & Engineering Geology, 2012, 39(1): 114-119. (in Chinese)
[2] 杜国梁, 杨志华, 袁颖, 等. 基于逻辑回归–信息量的川藏交通廊道滑坡易发性评价[J]. 水文地质工程地质, 2021, 48(5): 102-111
Du Guoliang, Yang Zhihua, Yuan Ying, et al. Landslide susceptibility mapping in the Sichuan-Tibet trafficcorridor using logistic regression- information value method[J]. Hydrogeology & Engineering Geology, 2021, 48(5): 102-111. (in Chinese)
[3] 黄发明, 胡松雁, 闫学涯, 等. 基于机器学习的滑坡易发性预测建模及其主控因子识别[J]. 地质科技通报, 2022, 41(2): 79–90
HUANG Faming, HU Songyan, YAN Xueya, et al. Landslide susceptibility prediction and identification of its main environmental factors based on machine learning models[J]. Bulletin of Geological Science and Technology 2022, 41(2): 79–90.
[4] 黄武彪, 丁明涛, 王栋, 等. 基于层数自适应加权卷积神经网络的川藏交通廊道沿线滑坡易发性评价[J]. 地球科学, 2022, 47(6): 2015-2030
Huang Wubiao, Ding Mingtao, Wang Dong, et al. Evaluation of landslide susceptibility based on layer adaptive weighted convolutional neural network model along Sichuan-tibet traffic torridor[J]. Earth Science, 2022, 47(6): 2015-2030. (in Chinese)
[5] 黄煜, 谢婉丽, 刘琦琦, 等. 基于GIS与MaxEnt模型的滑坡易发性评价—以铜川市中部城区为例[J]. 西北地质, 2023, 56(1): 266-275.
HUANG Yu, XIE Wanli, LIU Qiqi, et al. Landslide Susceptibility Assessment Based on GIS and MaxEnt Model: Example from Central Districts in Tongchuan City[J]. Northwestern Geology, 2023, 56(1): 266−275.
[6] 贾俊, 毛伊敏, 孟晓捷, 等. 深度随机森林和随机森林算法的滑坡易发性评价对比—以汉中市略阳县为例[J]. 西北地质, 2023, 56(3): 239−249.
JIA Jun, MAO Yimin, MENG Xiaojie, et al. Comparison of Landslide Susceptibility Evaluation by Deep Random Forest and Random Forest Model: A Case Study of Lueyang County, Hanzhong City[J]. Northwestern Geology, 2023, 56(3): 239−249.
[7] 康孟羽, 朱月琴, 陈晨, 等. 基于多元非线性回归和BP神经网络的滑坡滑动距离预测模型研究[J]. 地质通报, 2022, 41(12): 2281−2289.
KANG Mengyu, ZHU Yueqin, CHEN Chen, et al. Research on landslide sliding distance prediction model based on multiple nonlinear regression and BP neural network[J]. Geological Bulletin of China, 2022, 41(12): 2281−2289.
[8] 李东, 周可法, 孙卫东, 等. BP神经网络和SVM在矿山环境评价中的应用分析[J]. 干旱区地理, 2015, 38(1): 128-134
Li Dong, Zhou Kefa, Sun WeiDong, et al. Application of BP neural network and SVM in mine environmental assessment[J]. Arid Land Geography, 2015, 38(1): 128-134. (in Chinese)
[9] 李宇嘉, 陈宁生, 杨溢, 等. 汉源县中海村“8·21”雨后型滑坡特征与成因[J]. 成都理工大学学报(自然科学版), 2022, 49(2): 185−195.
LI Ningjia, CHENG Ningsheng, YANG Yi, et al. Characteristics and causes of 8·21 rainfall-induced landslide at Zhonghai Village, Hanyuan County, Sichuan, China[J]. Journal of Chengdu University of Technology (Science & Technology Edition), 2022, 49(2): 185−195.
[10] 林荣福, 刘纪平, 徐胜华, 等. 随机森林赋权信息量的滑坡易发性评价方法[J]. 测绘科学, 2020, 45(12): 131-138
Lin Rongfu, Liu Jiping, Xu Shenghua, et al. Evaluation method of landslide susceptibility based on random forest weighted information[J]. Science of Surveying and Mapping, 2020, 45(12): 131-138. (in Chinese)
[11] 刘坚, 李树林, 陈涛. 基于优化随机森林模型的滑坡易发性评价[J]. 武汉大学学报(信息科学版), 2018, 43(7): 1085-1091
Liu Jian, Li Shulin, Chen Tao. Landslide susceptibility assesment based on optimized random forest model[J]. Geomatics and Information Science of Wuhan University, 2018, 43(7): 1085-1091. (in Chinese)
[12] 刘璐瑶, 高惠瑛, 李照. 基于CF与Logistic回归模型耦合的永嘉县滑坡易发性评价[J]. 中国海洋大学学报(自然科学版), 2021, 51(10): 121-129
Liu Luyao, Gao Huiying, Li Zhao. Landslide susceptibility assessment based on coupling of CF model and logistic regression model in Yongjia county[J]. Periodical of Ocean University of China, 2021, 51(10): 121-129. (in Chinese)
[13] 刘希林, 庙成, 田春山. 区域滑坡和泥石流灾害两种危险性评价方法的比较分析[J]. 防灾减灾工程学报, 2017, 37(1): 71-78
Liu Xilin, Miao Cheng, Tian Chunshan. Comparative analysis of two methods for assessing hazard of landslide and debris-flow on a regional scale[J]. Journal of Disaster Prevention and Mitigation Engineering, 2017, 37(1): 71-78. (in Chinese)
[14] 马啸, 王念秦, 李晓抗, 等. 基于RF-FR模型的滑坡易发性评价——以略阳县为例[J]. 西北地质, 2022, 55(3): 335−344.
MA Xiao, WANG Nianqin, LI Xiaokang, YAN Dong, LI Jialin. Assessment of Landslide Susceptibility Based on RF-FR Model: Taking Lueyang County as an Example[J]. Northwestern Geology, 2022, 55(3): 335−344.
[15] 孟晓捷, 张新社, 曾庆铭, 等. 基于加权信息量法的黄土滑坡易发性评价——以1: 5万天水市麦积幅为例[J]. 西北地质, 2022, 55(2): 249−259.
MENG Xiaojie, ZHANG Xinshe, ZENG Qingming, et al. The Susceptibility Evaluation of Loess Landslide Based on Weighted Information Value Method: Taking 1: 50 000 Map of Maiji District of Tianshui City As an Example[J]. Northwestern Geology, 2022, 55(2): 249−259.
[16] 唐睿旋, 晏鄂川, 唐薇. 基于粗糙集和BP神经网络的滑坡易发性评价[J]. 煤田地质与勘探, 2017, 45(6): 129-138
Tan Ruixuan, Yan Echuan, Tan Wei. Landslide susceptibility evaluation based on rough set and back-propagation neural network[J]. Coal Geology & Exploration, 2017, 45(6): 129-138. (in Chinese)
[17] 孙萍萍, 张茂省, 贾俊, 等. 中国西部黄土区地质灾害调查研究进展[J]. 西北地质, 2022, 55(3): 96−107.
SUN Pingping, ZHANG Maosheng, JIA Jun, et al. Geo-hazards Research and Investigation in the Loess Regions of Western China[J]. Northwestern Geology, 2022, 55(3): 96−107.
[18] 田媛, 巨能攀, 解明礼, 等. 滑坡编录表达模式对易发性评价结果的影响[J]. 成都理工大学学报(自然科学版), 2022, 49(5): 606−615.
TIAN Yuan, JU Nengpan, XIE Mingli, et al. Analysis of the impact of landslide cataloguing expression patterns on the evaluation results of landslide susceptibility[J], Journal of Chengdu University of Technology (Science Technology Edition), 2022, 49(5): 606−615.
[19] 王朋伟, 安玉科. 滑坡匀速变形阶段快速诊断方法研究[J]. 西北地质, 2023, 56(5): 197−203.
WANG Pengwei, AN Yuke. Research on Rapid Diagnosis Method of Landslide’s Uniform Deformation Stage[J]. Northwestern Geology, 2023, 56(5): 197−203.
[20] 王海芝, 王颂, 周剑, 等. 樟木堆积体斜坡动力稳定性与极限承载力评价[J]. 西北地质, 2022, 55(1): 262−273.
WANG Haizhi, WANG Song, ZHOU Jian, et al. Dynamic Stability Analysis and Ultimate Bearing Capacity Evaluation of Zhangmu Landslide Deposit[J]. Northwestern Geology, 2022, 55(1): 262−273.
[21] 王毅, 方志策, 牛瑞卿, 等. 基于深度学习的滑坡灾害易发性分析[J]. 地球信息科学学报, 2021, 23(12): 2244-2260 doi: 10.12082/dqxxkx.2021.210057
Wang Yi, Fang Zhice, Niu Ruiqing, et al. Landslide susceptibility analysis based on deep learning[J]. Journal of Geo-information Science, 2021, 23(12): 2244-2260. (in Chinese) doi: 10.12082/dqxxkx.2021.210057
[22] 吴孝情, 赖成光, 陈晓宏, 等. 基于随机森林权重的滑坡危险性评价: 以东江流域为例[J]. 自然灾害学报, 2017, 26(5): 119-129
Wu Xiaoqing, Lai Chengguang, Chen Xiaohong, et al. A landslide hazard assessment based on random forest weight: a case study in the Dongjiang River Basin[J]. Journal of Natural Disasters, 2017, 26(5): 119-129. (in Chinese)
[23] 杨平芬, 李艳美, 杨保住. 云南省芒市土壤肥力现状及变化趋势[J]. 云南农业科技, 2014, (5): 4-7 doi: 10.3969/j.issn.1000-0488.2014.05.001
Yang Pingfen, Li Yanmei, Yang Baozhu. Status and trends of soil fertility in Mangshi, Yunnan province[J]. Yunnan Agricultural Science and Technology, 2014, (5): 4-7. (in Chinese) doi: 10.3969/j.issn.1000-0488.2014.05.001
[24] 张林梵, 王佳运, 张茂省, 等. 基于BP神经网络的区域滑坡易发性评价[J]. 西北地质, 2022, 55(2): 260-270
Zhang Linfan, Wang Jiayun, Zhang Maosheng, et al. Evaluation of regional landslide susceptibility assessment based on BP neural network[J]. Northwestern Geology, 2022, 55(2): 260-270. (in Chinese)
[25] 郑迎凯, 陈建国, 王成彬, 等. 确定性系数与随机森林模型在云南芒市滑坡易发性评价中的应用[J]. 地质科技通报, 2020, 39(6): 131-144
Zheng Yingkai, Chen Jianguo, Wang Chengbin, et al. Application of certainty factor and random forests model in landslide susceptibility evaluation in Mangshi city[J]. Bulletin of Geological Science and Technology, 2020, 39(6): 131-144. (in Chinese)
[26] 周定义, 左小清, 喜文飞, 等. 联合SBAS-InSAR和PSO-BP算法的高山峡谷区地质灾害危险性评价[J]. 农业工程学报, 2021, 37(23): 108-116 doi: 10.11975/j.issn.1002-6819.2021.23.013
Zhou Dingyi, Zuo Xiaoqing, Xi Wenfei, et al. Combined SBAS-InSAR and PSO-BP algorithm for evaluating the risk of geological disasters in alpine valley regions[J]. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(23): 108-116. (in Chinese) doi: 10.11975/j.issn.1002-6819.2021.23.013
[27] 周宇, 常鸣, 孙文静, 等. 基于改进证据权重法的北海道地震同震滑坡易发性评价[J]. 地理与地理信息科学, 2022, 38(1): 138-144
Zhou Yu, Chang Ming, Sun Wenjing, et al. Susceptibility evaluation of Hokkaido earthquake coseismic landslides based on improved weights of evidence method[J]. Geography and Geo-Information Science, 2022, 38(1): 138-144. (in Chinese)
[28] Adnan Ozdemir, Tolga Altural. A comparative study of frequency ratio, weights of evidence and logistic regression methods for landslide susceptibility mapping: Sultan Mountains, SW Turkey[J]. Journal of Asian Earth Sciences, 2013, 64: 180-197. doi: 10.1016/j.jseaes.2012.12.014
[29] L Ayalew, H Yamagishi. The application of GIS-based logistic regression for landslide susceptibility mapping in the Kakuda-Yahiko Mountains, Central Japan[J]. Geomorphology, 2005, 65(1): 15-31.
[30] Binh Thai Pham, Biswajeet Pradhan, Dieu Tien Bui, et al. A comparative study of different machine learning methods for landslide susceptibility assessment: A case study of Uttarakhand area (India)[J]. Environmental Modelling & Software, 2016, 84: 240-250.
[31] Wei Chen, Hamid Reza Pourghasemi, Mahdi Panahi, et al. Spatial prediction of landslide susceptibility using an adaptive neuro-fuzzy inference system combined with frequency ratio, generalized additive model, and support vector machine techniques[J]. Geomorphology, 2017, 297: 69-85. doi: 10.1016/j.geomorph.2017.09.007
[32] FC Dai, CF Lee, YY Ngai. Landslide risk assessment and management: an overview[J]. Engineering Geology, 2002, 64(1): 65-87. doi: 10.1016/S0013-7952(01)00093-X
[33] Francisco Gutiérrez, Rogelio Linares, Carles Roqué, et al. Large landslides associated with a diapiric fold in Canelles Reservoir (Spanish Pyrenees): Detailed geological geomorphological mapping, trenching and electrical resistivity imaging[J]. Geomorphology, 2015, 241: 224-242. doi: 10.1016/j.geomorph.2015.04.016
[34] Harlow L L. The Essence of multivariate thinking: Basic themes and methods[M]. London: Lawrence ErlbaumAssociates, 2005: 199-207.
[35] Haydar Y. Hussin, Veronica Zumpano, Paola Reichenbach, et al. Different landslide sampling strategies in a grid-based bi-variate statistical susceptibility model[J]. Geomorphology, 2016, 253: 508-523. doi: 10.1016/j.geomorph.2015.10.030
[36] Rumelhart d Hinton-G-Williams-R. . Learning representations by back-propagatingerrors[J]. Nature, 1986, 323: 533-536. doi: 10.1038/323533a0
[37] Faming Huang, Zhongshan Cao, Jianfei Guo, et al. Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping[J]. Catena, 2020, 191: 104580. doi: 10.1016/j.catena.2020.104580
[38] Guan-Wei Lin, Hongey Chen, Tsun-Ying Shih, et al. Various links between landslide debris and sediment flux during earthquake and rainstorm events[J]. Journal of Asian Earth Sciences, 2012, 54-55: 41-48. doi: 10.1016/j.jseaes.2012.03.012
[39] Markus Meinhardt, Manfred Fink, Hannes Tuenschel. Landslide susceptibility analysis in central Vietnam based on an incomplete landslide inventory: Comparison of a new method to calculate weighting factors by means of bivariate statistics[J]. Geomorphology, 2015, 234: 80-97. doi: 10.1016/j.geomorph.2014.12.042
[40] Pham, BT, Shirzadi, et al. A hybrid machine learning ensemble approach based on a Radial Basis Function neural network and Rotation Forest for landslide susceptibility modeling: A case study in the Himalayan area, India[J]. Int J Sediment Res, 2018, 2018, 33(2): 157-170.
[41] B. Pradhan. Landslide susceptibility mapping of a catchment area using frequency ratio, fuzzy logic and multivariate logistic regression approaches[J]. Journal of the Indian Society of Remote Sensing, 2010, 38(2): 301-320. doi: 10.1007/s12524-010-0020-z
[42] Ryuta Tamura, Ken Kobayashi, Yuichi Takano, et al. Mixed integer quadratic optimization formulations for eliminating multicollinearity based on variance inflation factor[J]. Journal of Global Optimization, 2019, 73(2): 431-446. doi: 10.1007/s10898-018-0713-3
[43] Shen Chaoyong, Feng Zhongke, Xie Chou, et al. Refinement of Landslide Susceptibility Map Using Persistent Scatterer Interferometry in Areas of Intense Mining Activities in the Karst Region of Southwest China[J]. Remote Sensing, 2019, 11(23): 2821-2843.
[44] Vijendra Kumar Pandey, Kaushal Kumar Sharma, Hamid Reza Pourghasemi, et al. Sedimentological characteristics and application of machine learning techniques for landslide susceptibility modelling along the highway corridor Nahan to Rajgarh (Himachal Pradesh), India[J]. Catena, 2019, 182: 104150. doi: 10.1016/j.catena.2019.104150
[45] Chao Zhou, Kunlong Yin, Ying Cao, et al. Application of time series analysis and PSO-SVM model in predicting the Bazimen landslide in the Three Gorges Reservoir, China[J]. Engineering Geology, 2016, 204: 108-120. doi: 10.1016/j.enggeo.2016.02.009
[46] Zhifu Zhu, Shu Gan, Xiping Yuan, et al. Landslide Susceptibility Mapping with Integrated SBAS-InSAR Technique: A Case Study of Dongchuan District, Yunnan (China)[J]. Sensors, 2022, 22(15): 5587. doi: 10.3390/s22155587
-