摘要
应用神经网络集成模型,以空气湿度、温度、太阳辐射以及风速为输入,利用交叉验证方法确定网络隐层节点数,建立作物需水量的预测模型。实验结果表明,与单个神经网络与随机森林模型相比,神经网络集成模型能获得更好的预测精度,可用于节水灌溉。
In order to predict crop water requirements, a method based on neural network ensemble is developed in this paper. Solar radiation, air temperature, relative humidity and wind speed are used as input variables. The number of hidden-layer neurons is determined by cross validation method. Experimental results indicate that the method based on neural network ensemble outperforms those based on a single neural network and random forest, and can be used for water-saving irrigation.
出处
《软件导刊》
2018年第1期46-48,共3页
Software Guide
基金
江苏省新型环保重点实验室项目(AE201121)
盐城市农业科技指导性计划项目(YKN2014012)
关键词
神经网络
集成模型
作物需水量
随机森林
neural network
ensemble model
crop water requirements
random forest
