摘要
参考作物腾发量(ET0)是估算作物腾发量的关键参数,其准确预测对提高作物需水预报精度具有十分重要的意义。Elman神经网络是BP网络的改进结构,具有适应时变性的特点;最小二乘支持向量机(LS-SVM)是支持向量机(SVM)的一种优化算法,它基于结构风险最小化准则,可兼顾模型的经验风险和推广能力。将两种方法应用于参考作物腾发量预测中,并以铁岭市为例,对比分析LS-SVM模型与Elman模型的预测值。结果表明:LS-SVM模型学习速度快,具有比Elman模型更高的模拟性能和预测精度,更适合参考作物腾发量的预测。
Reference evaportranspiration is a key parameter in estimating crop evaportranspiration.Its prediction is very important in estimating crop evapotranspiration and in improving the using efficiency of agricultural water.The Elman neural network is a dynamic neural network based on BP neural network.LS-SVM is an improvement of SVM algorithm.It is based on the minimum structure risk,which can give dual attention to the experience risk of the model and promoted ability.Elman and LS-SVM was used in reference evaportranspiration prediction.A comparison and analysis of prediction by the two models was given in the paper.The results show that the LS-SVM model not only has a quicker study speed,but also has a better predictable precision and stability.
出处
《农业科技与装备》
2011年第6期88-90,共3页
Agricultural Science & Technology and Equipment
