| 1,729 | 3 | 916 |
| 下载次数 | 被引频次 | 阅读次数 |
【目的】洪水是由暴雨、冰雪快速融化、风暴潮等因素引发的自然灾害，常导致经济损失与生活不便。常规的洪水预测主要依赖于传统的水文学方法和基于经验的统计模型，但在遇到缺乏长期、连续的水文观测数据的地区，利用其他数据进行洪水预测的方法就至关重要。【方法】基于决策树的机器学习算法，如随机森林、XGBoost及LightGBM,因其直观性和强大功能，在分类和回归任务中具有良好的预测能力，适用于洪水预测。使用包含50 000条记录与21个变量的数据集，评估随机森林、XGBoost和LightGBM三种算法的洪水预测能力，通过预测效果与关键变量识别比较其性能，并以ROC-AUC曲线衡量优劣。【结果】结果显示：所有模型均表现出较高的预测精度，其中XGBoost模型具有最小的均方误差0.000 186 2和最高的决定系数0.925 2,而LightGBM模型在ROC-AUC曲线中取得了最大的AUC值0.99。随机森林模型各指标均不如以上二者。【结论】结果表明：XGBoost模型在洪水概率的预测方面效果最好，预测误差最小；而对于预测洪水是否发生这类二分类情况，LightGBM则是最优的选择。
Abstract：[Objective]Floods are natural disasters triggered by factors such as heavy rainfall, rapid snow and ice melt, and storm surges, often resulting in significant economic losses and severe disruption to daily life. Conventional flood prediction primarily relies on traditional hydrological method and experience-based statistical models. However, in areas lacking long-term and continuous hydrological monitoring data, alternative data-driven method for flood prediction are essential.[Methods]Machine learning algorithms based on decision trees, including Random Forest, XGBoost, and LightGBM, demonstrated excellent performance in classification and regression tasks due to their interpretability and strong functions, making them suitable for flood prediction. A dataset containing 50 000 records and 21 variables was used to evaluate the flood prediction performance of these three algorithms, namely Random Forest, XGBoost, and LightGBM. Their performance was assessed based on prediction accuracy and key variable identification, with the ROC-AUC curve used for comparative analysis.[Results]The result showed that all three models achieved high prediction accuracy. Among them, the XGBoost model exhibited the lowest mean squared error(0.000 186 2) and the highest coefficient of determination(0.925 2). Moreover, the LightGBM model achieved the highest AUC value(0.99) in the ROC-AUC curve. The Random Forest model underperformed the other two across all indicators.[Conclusion]The findings indicate that XGBoost delivers optimal performance for flood probability prediction with lowest prediction errors, while LightGBM is the optimal choice for binary classification tasks, such as predicting flood occurrence.
[1] 李国英.为以中国式现代化全面推进强国建设、民族复兴伟业提供有力的水安全保障：在2024年全国水利工作会议上的讲话[J].水利发展研究，2024,24(1):1-10.LI G Y.Improved water security for China’s efforts to build itself into a stronger country and rejuvenate the Chinese nation on all fronts by pursuing Chinese modernization:Speech at the 2024 National Water Conservancy Work Conference[J].Water Resources Development Research,2024,24(1):1-10.
[2] EL-BAGOURY H,GAD A.Integrated hydrological modeling for watershed analysis,flood prediction,and mitigation using meteorological and morphometric data,SCS-CN,HEC-HMS/RAS,and QGIS[J].Water,2024,16(2):356.
[3] MURPHY K P.Probabilistic machine learning:An introduction[M].Cambridge:MIT press,2022.
[4] UTOMO D,HU L C,HSIUNG P A.Deep neural network-based spatiotemporal heterogeneous data reconstruction for landslide detection[J].International Journal of Data Science and Analytics,2024,17(1):93-109.
[5] SUN F Y,HAO W N,ZOU A,et al.A survey on spatio-temporal series prediction with deep learning:Taxonomy,applications,and future directions[J].Neural Computing and Applications,2024,36(17):9919-9943.
[6] HAN D,CHAN L,ZHU N.Flood forecasting using support vector machines [J].Journal of Hydroinformatics,2007,9 (4):267-276.
[7] AL SAWAF M B,KAWANISI K,JLILATI M N,et al.Extent of detection of hidden relationships among different hydrological variables during floods using data-driven models[J].Environmental Monitoring and Assessment,2021,193(11):692.
[8] COSTA V G,PEDREIRA C E.Recent advances in decision trees:an updated survey[J].Artificial Intelligence Review,2023,56(5):4765-4800.
[9] SAHOO A,GHOSE D K.Flood Forecasting Using Hybrid SVM-GOA Model:A Case Study[M].Singapore:Springer Nature Singapore,2022:407-416.
[10] AGARWAL S,ROY P J,CHOUDHURY P,et al.Flood forecasting and flood flow modeling in a river system using ANN[J].Water Practice and Technology,2021,16(4):1194-1205.
[11] LAWAL Z K,YASSIN H,ZAKARI R Y.Flood prediction using machine learning models:A case study of kebbi state Nigeria[C]//IEEE.2021 IEEE Asia-Pacific Conference on Computer Science and Data Engineering (CSDE).Brisbane,Australia,2021:1-6.
[12] SYEED M M A,FARZANA M,NAMIR I,et al.Flood prediction using machine learning models[C]//IEEE.2022 International Congress on Human-Computer Interaction,Optimization and Robotic Applications (HORA).Ankara,Turkey,2022:1-6.
[13] GRZESIAK M,THAKKAR P.Flood prediction using classical and quantum machine learning models[EB/OL].2024:2407.01001.https://arxiv.org/abs/2407.01001v1.
[14] BOATENG E Y,OTOO J,ABAYE D A.Basic tenets of classification algorithms K-nearest-neighbor,support vector machine,random forest and neural network:A review[J].Journal of Data Analysis and Information Processing,2020,8(4):341-357.
[15] SCHOPPA L,DISSE M,BACHMAIR S.Evaluating the performance of random forest for large-scale flood discharge simulation[J].Journal of Hydrology,2020,590:125531.
[16] SAGI O,ROKACH L.Approximating XGBoost with an interpretable decision tree[J].Information Sciences,2021,572:522-542.
[17] AHN J M,KIM J,KIM K.Ensemble machine learning of gradient boosting (XGBoost,LightGBM,CatBoost) and attention-based CNN-LSTM for harmful algal blooms forecasting[J].Toxins,2023,15(10):608.
[18] SAHASTRABUDDHE R,GHAUSI S A,JOSEPH J,et al.Indian summer monsoon rainfall in a changing climate:A review[J].Journal of Water and Climate Change,2023,14(4):1061-1088.
[19] GOSWAMI B N,VENUGOPAL V,SENGUPTA D,et al.Increasing trend of extreme rain events over India in a warming environment[J].Science,2006,314(5804):1442-1445.
[20] FLEISCHMANN A,PAIVA R,COLLISCHONN W.Can regional to continental river hydrodynamic models be locally relevant?A cross-scale comparison[J].Journal of Hydrology X,2019,3:100027.
[21] POFF N L,ALLAN J D,BAIN M B,et al.A paradigm for river conservation and restoration[J].BioScience,1997,47(11):769-784.
[22] JACOBSON C R.Identification and quantification of the hydrological impacts of imperviousness in urban catchments:a review[J].Journal of Environmental Management,2011,92(6):1438-1448.
[23] Intergovernmental Panel on Climate Change (IPCC).Climate Change 2014:Impacts,Adaptation and Vulnerability:Part A:Global and Sectoral Aspects[M].Cambridge,UK:Cambridge University Press,2014.
[24] WALLING D E.Human impact on land-ocean sediment transfer by the world’s rivers[J].Geomorphology,2006,79(3/4):192-216.
[25] PIMENTEL D,HARVEY C,RESOSUDARMO P,et al.Environmental and economic costs of soil erosion and conservation benefits[J].Science,1995,267(5201):1117-1123.
[26] TIERNEY K J.Recent Developments in U.S.Homeland Security Policies and Their Implications for the Management of extreme Events[M].New York,NY:Springer New York,2007:405-412.
[27] 米胤瑜，孔锋.气候变化背景下城市洪水风险管理体系国际比较与启示：以伦敦、纽约、郑州为例[J].水利水电技术(中英文),2023,54(3):21-34.MI Y Y,KONG F.International comparison and enlightenment of urban flood risk management system under background of climate change:Taking London,New York and Zhengzhou as study cases[J].Water Resources and Hydropower Engineering,2023,54(3):21-34.
[28] WONG P P,LOSADA I J,GATTUSO J P,et al.Coastal systems and low-lying areas[J].Climate Change,2014,2104:361-409.
[29] WAGENER T,SIVAPALAN M,TROCH P,et al.Catchment classification and hydrologic similarity[J].Geography Compass,2007,1(4):901-931.
[30] 张海凤，孔锋，方建.超常规极端暴雨洪涝灾害应对的国际比较研究：以 2021 年中美德暴雨洪涝灾害为例[J] .水利水电技术 (中英文),2023,54(7):1-13.ZHANG Haifeng,KONG Feng,FANG Jian.International comparative study on coping with flood-waterlogging disaster from extraordinary rainstorm:Taking rainstorm flood-waterlogging disasters in China,America and Germany in 2021 as study cases[J] .Water Resources and Hydropower Engineering,2023,54(7):1-13.
[31] MCGRANAHAN G,BALK D,ANDERSON B.The rising tide:Assessing the risks of climate change and human settlements in low elevation coastal zones[J].Environment and Urbanization,2007,19(1):17-37.
[32] ZEDLER J B,KERCHER S.Wetland Resources:status,trends,ecosystem services,and restorability[J].Annual Review of Environment and Resources,2005,30:39-74.
[33] BREIMAN L.Random forests[J].Machine learning,2001,45:5-32.
[34] LIAW A,WIENER M.Classification and regression by randomForest[J].R News,2002,2(3):18-22.
[35] GUOLINKE Q M,FINLEY T,WANG T,et al.Lightgbm:A highly efficient gradient boosting decision tree[J].Advances in Neural Information Rocessing Systems,2017,30:52.
[36] CHEN T Q,GUESTRIN C.XGBoost:A scalable tree boosting system[C]//ACM.Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.San Francisco California USA:ACM,2016:785-794.
[37] 李红艳，郝景开，刘大为，等.基于元启发式算法优化的洪水风险评价模型[J].水资源保护，2024,40(6):1-15.LI Hongyan,HAO Jingkai,LIU Dawei,et al.Flood risk assessment model optimized based on meta-heuristic algorithm[J].Water Resources Protection,2024,40(6):1-15.
[38] 桑珍珍，崔杰，闫寒，等.基于随机森林和XGBoost算法构建心脏骤停患者自主循环恢复后神经功能预后不良的风险预测模型[J].中国急救医学，2024,44(7):577-585.SANG Z Z,CUI J,YAN H,et al.Construction of prediction model for poor prognosis of neurological function in cardiac arrest patients after the return of spontaneous circulation based on Random Forest and XGBoost algorithms[J].Chinese Journal of Critical Care Medicine,2024,44(7):577-585.
[39] VINGE R,MCKELVEY T.Understanding support vector machines with polynomial kernels[C]//IEEE.2019 27th European Signal Processing Conference (EUSIPCO).Spain:A Coruna,2019:1-5.
[40] BUHRMESTER V,MüNCH D,ARENS M.Analysis of explainers of black box deep neural networks for computer vision:A survey[J].Machine Learning and Knowledge Extraction,2021,3(4):966-989.
[41] LECUN Y,BENGIO Y,HINTON G.Deep learning[J].Nature,2015,521(7553):436-444.
[42] DIETRICH R,OPPER M,SOMPOLINSKY H.Statistical mechanics of support vector networks[J].Physical Review Letters,1999,82(14):2975-2978.
[43] ALI M,TAHA M,AZIZ M S,et al.Flash flood prediction of panjkora river,kpk,using artificial neural networks (ANN) and support vector machine (SVM)[J].Technical Journal,2024,3(ICACEE):758-769.
基本信息:
DOI：10.13928/j.cnki.wrahe.2025.08.002
中图分类号:P426.616;TV87
引用信息:
[1]左右.基于决策树的水灾预测模型[J].水利水电技术(中英文),2025,56(08):19-31.DOI:10.13928/j.cnki.wrahe.2025.08.002.
基金信息:
国家社会科学基金项目(21BTJ071)
2024-10-15
2024-10-15
2024-10-15