摘要
科学估算地下水资源量,对合理开发利用地下水,保护生态及地质环境,促进经济社会和生态环境协调发展具有积极意义。统计地下水资源是一项复杂而困难的工作,而地下水资源量与降水量和地表水资源关系密切,但它们之间的关系十分复杂,难以用数学解析式准确表达。随机森林通过某种机制组合多个弱的决策树,最终使输出决策结果具有较高的精确。运用随机森林模型构建了湖北省地下资源与降水量和地表水资源之间的非线性映射关系,通过分析降水量和地表水资源数据,可达到预测地表水资源的目的。结果显示,平均预测误差仅1.1522%,比BP神经网络的2.1111%减小了45.4218%。
scientific estimation of groundwater resources is of positive significance for the rational development and utilization of groundwater,protection of ecological and geological environment,and promotion of coordinated development of economic society and ecological environment.Statistics of groundwater resources is a complex and difficult work,and the groundwater resources are closely related to precipitation and surface water resources,but the relationship between them is very complex,and it is difficult to express accurately by mathematical analysis.Random forest,combining several weak decision trees through some mechanism,finally attains the output decision results with higher accuracy.The non-linear mapping relationship between groundwater resources,precipitation and surface water resources in Hubei Province is constructed by using random forest model.The purpose of predicting surface water resources can be achieved by analyzing the data of precipitation and surface water resources.The results show that the average prediction error is only 1.1522%,which is 45.4218%less than the 2.1111%of BP neural network.
作者
舒服华
SHU FU HUA(School of mechanical and electrical engineering,Wuhan University of technology,Wuhan,Hubei 430070)
出处
《武汉电力职业技术学院学报》
2020年第2期66-70,共5页
Journal of Wuhan Electric Power Technical College
关键词
湖北
地下水资源
预测
降水量
地表水资源
随机森林
Hubei
Groundwater resources
Prediction
Precipitation
Surface water resources
Random forest