基于随机森林模型的近岸海域水质污染预测预警方法研究
Study on Water Pollution Prediction and Early Warning Method in Coastal Water Based on Random Forest Model
罗俊 1李博阳 1黄琳 2母凌燕1
作者信息
- 1. 深圳市衡兴安全检测技术有限公司,广东深圳 518048
- 2. 清华大学深圳国际研究生院,广东深圳 518000
- 折叠
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.
Abstract
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.
关键词
随机森林模型/近岸海域/水质污染/预测/预警方法Key words
Random forest model/Nearshore waters/Water pollution/Prediction/Early warning method引用本文复制引用
基金项目
2021年深圳市可持续发展专项课题(KCXFZ20201221173401005)
出版年
2024
NETL
NSTL
万方数据