摘要
常规的近岸海域水质污染预测预警方法以水环境监测为主,受到温度、湿度、浊度等水质参数的影响,污染预测预警不准确将影响近岸水域水质保护。设计了基于随机森林模型的近岸海域水质污染预测预警方法,采集并预处理近岸海域水质数据,基于随机森林模型预测近岸海域水质污染物的空间布局,将水质污染物数量作为随机森林树的数量,再将污染预警等级作为树的深度,通过交叉验证的方式优化随机森林模型参数,满足水质污染预测需求,将近岸海域水质污染物浓度预测值作为预警阈值,划分出近岸海域水质污染预警等级,确保预警效果。通过对比实验,验证了该方法的预测预警效果更加准确,能够应用于实际生活中。
The conventional prediction and early warning method for water pollution in coastal water is mainly water environment monitoring,which is affected by water quality parameters,such as temperature,humidity and turbidity,etc.The inaccurate pollution prediction and early warning affects the water quality protection in coastal water.The study designs coastal water pollution prediction and early warning method based on random forest model,collects and preprocesses water quality data from nearshore areas,predicts the spatial layout of water pollutants in nearshore areas based on random forest model,uses the number of water pollutants as the number of random forest trees,and then uses the pollution warning level as the depth of the trees.Through cross validation,the study optimizes the parameters of the random forest model to meet the needs of water pollution prediction;divides the water quality pollution warning levels in nearshore areas,uses the predicted concentration of water pollutants in nearshore areas as the warning threshold,and classifies the water quality pollution warning levels in nearshore areas to ensure the warning effect.Through comparative experiments,it is verified that the prediction and early warning effect of this method is more accurate,and can be applied in practical life.
作者
罗俊
李博阳
黄琳
母凌燕
Luo Jun;Li Boyang;Huang Lin;Mu Lingyan(Shenzhen Hengxing Safety Testing Technology Co.,LTD,Shenzhen 518048,China;Tsinghua University Shenzhen International Graduate School,Shenzhen 518000,China)
出处
《黑龙江科学》
2024年第12期22-25,共4页
Heilongjiang Science
基金
2021年深圳市可持续发展专项课题“专2021N017近岸海域环境风险监测系统与预警技术研发”(KCXFZ20201221173401005)。
关键词
随机森林模型
近岸海域
水质污染
预测
预警方法
Random forest model
Nearshore waters
Water pollution
Prediction
Early warning method
