基于随机森林算法的洪水灾害风险评估研究
|
摘要
根据流域灾害系统理论,在考虑致灾因子、孕灾环境和承灾体的基础上,选取9个风险评价指标,运用样本数据进行人工识别风险并得到训练样本,采用随机森林算法构建基于随机森林的洪水灾害风险评估模型。然后采用随机森林自评估工具,分析建立的洪水灾害风险评估模型的误差和指标,同时构建支持向量机模型作为对比方案,并采用五折交叉验证方法对基于随机森林算法的洪水灾害风险评估模型和支持向量机模型进行验证。最后以海河流域邱庄段为研究对象,分别运用基于随机森林算法的洪水灾害风险评估模型和基于支持向量机模型对相同的数据集进行评估和对比,结果显示,12 h内降雨总量、洪水持续时间和土壤含水量是引发洪水的主要因素,而基于随机森林算法的洪水灾害风险评估的训练精度及测试精度均高于支持向量机模型。
According to the theory of catchment disaster system,firstly 9 risk evaluation indices are selected considering disastercausing factors, disaster-pregnant environment and disasterbearing bodies,the sample data are used to manually identify risks and to obtain training samples,and a risk assessment model for flood disasters is established by using the random forest algorithm based on the random forests. Secondly, the selfassessment tools of the random forests are employed to analyze the model errors and indexes,and the support vector machine model is established as a comparison scheme. The 5-fold crossvalidation method is used to validate the above two models.Finally,the Qiuzhuang catchment of Haihe River Basin is taken as the research object to evaluate and compare the evaluation models based on the random forests and the support vector machine with the same data. The results show that the total rainfall in 12 hours,duration of floods and water content of soils are the main factors to cause floods. The training and testing accuracies of the risk assessment of flood disasters based on the random forests are both higher than those of the support vector machine.
According to the theory of catchment disaster system,firstly 9 risk evaluation indices are selected considering disastercausing factors, disaster-pregnant environment and disasterbearing bodies,the sample data are used to manually identify risks and to obtain training samples,and a risk assessment model for flood disasters is established by using the random forest algorithm based on the random forests. Secondly, the selfassessment tools of the random forests are employed to analyze the model errors and indexes,and the support vector machine model is established as a comparison scheme. The 5-fold crossvalidation method is used to validate the above two models.Finally,the Qiuzhuang catchment of Haihe River Basin is taken as the research object to evaluate and compare the evaluation models based on the random forests and the support vector machine with the same data. The results show that the total rainfall in 12 hours,duration of floods and water content of soils are the main factors to cause floods. The training and testing accuracies of the risk assessment of flood disasters based on the random forests are both higher than those of the support vector machine.
引文
[1] 程晓陶.城市型水灾害及其综合治水方略[J].灾害学,2010(25):10-15.
[2] 张琳,孔小玲.关于洪水风险的可保性分析[J].统计与决策,2009(14):51-53.
[3] 冯婧,严登华,秦天玲,等.变化环境下我国西北内陆河区洪灾综合应对[J].水资源保护,2012,28(4):17-23.
[4] 韩平,程先富.洪水灾害损失评估研究综述[J].环境科学与管理,2012,37(4):61-64.
[5] 张行南,罗健,陈雷,等.中国洪水灾害危险程度区划[J].水利学报,2003(3):1-7.
[6] 周寅康.流域性洪涝及其指标研究[J].灾害学,1995,10(3):6-10.
[7] 曹永强,杜国志,王方雄.洪灾损失评估方法及其应用研究[J].辽宁师范大学学报(自然科学版),2006,29(3):355-358.
[8] STANLEY A,CHANGNON J.Research agenda for floods to solve policy failure[J].Journal of Water Resources Planning and Management,1985,111(1):54-64.
[9] EL JABI,N,ROUSSELLE J.A flood damage model for flood plain studies[J].Journal of the American Water Resources Association,1987,23(2):179-187.
[10] 程晓陶.加强水旱灾害管理的战略需求与治水方略的探讨[J].水利学报,2008,39(10):1197-1203.
[11] 王慧敏,刘高峰,佟金萍,等.非常规突发水灾害事件动态应急决策模式探讨[J].软科学,2012(1):20-24.
[12] 黄崇福.自然灾害风险评价理论与实践[M].北京:科学出版社,2004.
[13] 王新洲,史文中,王树良.模糊空间信息处理[M].武汉:武汉大学出版社,2003.
[14] 屈艳萍,高辉,吕娟,等.基于区域灾害系统论的中国农业旱灾风险评估[J].水利学报,2015(8):908-917.
[15] BREIMAN L.Random forests[J].Machine Learning,2001,45(1):5-32.
[16] 史培军.再论灾害研究的理论与实践[J].自然灾害学报,1996,5(4):6-17.
[17] United Nations Department of Humanitarian Affairs (UNDHA).Mitigating natural disasters:phenomena,effects and options:A manual for policy makers and planners[M].New York:United Nations,1991:1-164.
[18] 李谢辉,韩荟芬.河南省黄河中下游地区洪灾损失评估与预测[J].灾害学,2014(1):87-92.
[19] 史培军.四论灾害系统研究的理论与实践[J].自然灾害学报,2005,14(6):1-7.
[20] 史培军.三论灾害研究的理论与实践[J].自然灾害学报,2002,11(3):1-9.
[21] BURTON I,KATERSR W.The environment as hazard[M].2nd.New York:The Guilford Press,1993.
[22] 蒋卫国,李京,陈云浩,等.区域洪水灾害风险评估体系(Ⅰ):原理与方法[J].自然灾害学报,2008(6):56-59.
[2] 张琳,孔小玲.关于洪水风险的可保性分析[J].统计与决策,2009(14):51-53.
[3] 冯婧,严登华,秦天玲,等.变化环境下我国西北内陆河区洪灾综合应对[J].水资源保护,2012,28(4):17-23.
[4] 韩平,程先富.洪水灾害损失评估研究综述[J].环境科学与管理,2012,37(4):61-64.
[5] 张行南,罗健,陈雷,等.中国洪水灾害危险程度区划[J].水利学报,2003(3):1-7.
[6] 周寅康.流域性洪涝及其指标研究[J].灾害学,1995,10(3):6-10.
[7] 曹永强,杜国志,王方雄.洪灾损失评估方法及其应用研究[J].辽宁师范大学学报(自然科学版),2006,29(3):355-358.
[8] STANLEY A,CHANGNON J.Research agenda for floods to solve policy failure[J].Journal of Water Resources Planning and Management,1985,111(1):54-64.
[9] EL JABI,N,ROUSSELLE J.A flood damage model for flood plain studies[J].Journal of the American Water Resources Association,1987,23(2):179-187.
[10] 程晓陶.加强水旱灾害管理的战略需求与治水方略的探讨[J].水利学报,2008,39(10):1197-1203.
[11] 王慧敏,刘高峰,佟金萍,等.非常规突发水灾害事件动态应急决策模式探讨[J].软科学,2012(1):20-24.
[12] 黄崇福.自然灾害风险评价理论与实践[M].北京:科学出版社,2004.
[13] 王新洲,史文中,王树良.模糊空间信息处理[M].武汉:武汉大学出版社,2003.
[14] 屈艳萍,高辉,吕娟,等.基于区域灾害系统论的中国农业旱灾风险评估[J].水利学报,2015(8):908-917.
[15] BREIMAN L.Random forests[J].Machine Learning,2001,45(1):5-32.
[16] 史培军.再论灾害研究的理论与实践[J].自然灾害学报,1996,5(4):6-17.
[17] United Nations Department of Humanitarian Affairs (UNDHA).Mitigating natural disasters:phenomena,effects and options:A manual for policy makers and planners[M].New York:United Nations,1991:1-164.
[18] 李谢辉,韩荟芬.河南省黄河中下游地区洪灾损失评估与预测[J].灾害学,2014(1):87-92.
[19] 史培军.四论灾害系统研究的理论与实践[J].自然灾害学报,2005,14(6):1-7.
[20] 史培军.三论灾害研究的理论与实践[J].自然灾害学报,2002,11(3):1-9.
[21] BURTON I,KATERSR W.The environment as hazard[M].2nd.New York:The Guilford Press,1993.
[22] 蒋卫国,李京,陈云浩,等.区域洪水灾害风险评估体系(Ⅰ):原理与方法[J].自然灾害学报,2008(6):56-59.