基于大数据挖掘的洪水灾害易发性评价系统设计  

Design of flood disaster susceptibility evaluation system based on big data mining

作　　者：曹贤龙 CAO Xianlong(ZhengZhou Business University, GongYi 451200,China)

机构地区：[1]郑州商学院,河南巩义451200

出　　处：《灾害学》2021年第3期10-14,共5页Journal of Catastrophology

基　　金：河南省科技厅科技攻关项目(182102210134);河南省教育厅人文社科项目(2019-ZZJH-185)。

摘　　要：针对目前在设计洪水灾害易发性评价系统时由于没有分析洪水灾害数据之间的关联性,从而导致洪水灾害易发性评价准确率较差的问题,提出基于大数据挖掘的洪水灾害易发性评价系统设计。首先设定系统的整体框架,并对其框架中的功能模块进行具体描述,再将获取的洪水灾害信息数据进行离散化计算,构建洪水灾害的数据训练集与数据验证集,将构建的数据集输入系统中利用关联分析法对数据进行关联性计算并对计算结果进行排列,以此完成对洪水灾害的易发性评价。检测结果表明,该方法设计的系统在对洪水灾害进行易发性评价时Kappa系数、正确率及AUC的值均呈上升趋势,表明该系统对洪水灾害易发性评价具有一定的可行性。In order to solve the problem of poor evaluation accuracy of flood hazard vulnerability due to the failure to analyze the correlation between flood hazard data when designing flood hazard vulnerability evaluation system,a flood hazard vulnerability evaluation system design based on big data mining was proposed.First set system's overall framework,and the description framework of functional modules,and then will get information about flood disaster data discretization calculation,construction of flood disaster data training set and validation set,will build a set of data input in the system using the correlation analysis on the data correlation calculation and the calculation results are arranged,in order to complete the liability of the flood disaster evaluation.The test results show that the Kappa coefficient,accuracy rate and AUC of the system designed by this method are on the rise when evaluating the flood disaster vulnerability,which indicates that the system is feasible to evaluate the flood disaster vulnerability to a certain extent.

关 键 词：洪水灾害 易发性评价 关联分析法 数据预处理 系统设计 

分 类 号：P333[天文地球—水文科学] X43[水利工程—水文学及水资源]