摘要
结合门限自回归模型与人工神经网络模型的建模思想 ,首次提出这两种方法的耦合模型 ,即门限人工神经网络模型。新模型的实质是一种分段非线性化的处理方法 ,是对现有门限模型分段线性化的很好改进。实例计算结果说明 ,新模型在洪水的预报中是有效的 。
According to the modeling concept of the threshold auto-regressive model and artificial neural networks model , a new method called threshold artificial neural networks , is firstly presented in the paper .Compared with the existing threshold model which is linear in each subsection , the essence of the new method is non-linear in each subsection . The result of calculation example shows the flood forecasting with this new method is efficient in flood forecasting and is universal and practical in predicting nonlinear dynamic time series.
出处
《四川水力发电》
2002年第4期34-36,共3页
Sichuan Hydropower
基金
国家自然科学基金重大项目资助 (项目编号 :5 0 0 9962 0 )
