摘要
吉林省永吉县存在大量的斜坡地质灾害,为了给永吉县斜坡地质灾害的防治和预警提供高效直观的分析模型,将吉林省永吉县作为研究区,选取高程、坡度、坡向、剖面曲率、平面曲率、距断层距离、岩性、距河流距离、年均降雨量、地形湿度指数和植被覆盖指数等11个评价因子,利用神经网络模型进行区域斜坡地质灾害易发性分析,再选用频率比、支持向量机模型进行对比。利用ROC曲线对模型的准确性进行验证分析,得出神经网络、频率比和支持向量机模型的成功率分别是91.3%、89.3%、90.2%,预测率分别是87.3%、84.3%、85.6%。结果表明:神经网络模型的精度最高,更适用于永吉县斜坡地质灾害的易发性评价。
There are large number of slope geological hazards in Yongji County of Jilin Province.In order to provide an efficient and intuitive analysis model for the prevention and warning of slope geological hazards,Yongji County of Jilin Province is used as a research area.Eleven evaluation factors including elevation,slope,slope direction,profile curvature,plane curvature,distance from faults,lithology,distance from rivers,average annual rainfall,terrain humidity index,and vegetation cover index are selected.Neural network model is used to analyze the susceptibility of regional slope geological hazards,followed by the comparison with frequency ratio and support vector machine models.Finally,the accuracy of the model is verified and analyzed using the ROC curve.The success rates of the neural network,frequency ratio,and support vector machine models are 91.3%,89.3% and90.2%,and the prediction rates are 87.3%,84.3% and 85.6%,respectively.The results show that the neural network model has the highest accuracy and is more suitable for the assessment of slope geological hazards in Yongji County.
作者
刘飞
秦胜伍
乔双双
窦强
扈秀宇
LIU Fei;QIN Sheng-wu;QIAO Shuang-shuang;DOU Qiang;HU Xiu-yu(College of Construction Engineering,Jilin University,Changchun 130026,China)
出处
《世界地质》
CAS
2019年第4期1166-1176,共11页
World Geology
基金
吉林省科技发展计划项目(20190303103SF,20170101001JC)
关键词
斜坡地质灾害易发性评价
神经网络
频率比
支持向量机
永吉县
slope geological hazards susceptibility evaluation
neural network
frequency ratio
support vector machine
Yongji County
