摘要
针对天然水体水质预测问题,提出了基于灰色预测和神经网络的组合模型。利用不同时段的数据建立不同的灰色模型,将所得结果用神经网络模型组合。应用组合预测方法对北京密云水库水中DO值进行预测,并与单纯灰色和单纯神经网络模型比较。结果表明组合模型的预测值相对误差更小,精度更高。
Aiming at natural water quality forecast,a gray forecast model combined with artificial neural network model is developed. Different gray models are developed according to data in different periods.Then their results are combined based on a neural network model.Water quality of Miyun Reservoir is forecasted based on the combination forecast model,compared with simple grey and simple neural network ones.The results show that relative errors of the predictive values based on combination forecast model are much smaller.Thus the results are more reliable.
出处
《系统工程》
CSSCI
CSCD
北大核心
2011年第9期105-109,共5页
Systems Engineering
基金
国家自然科学基金资助项目(4107132271031001)
关键词
水质
灰色模型
神经网络
组合预测
Water Quality
Gray Model
Artificial Neural Networks
Combination Forecast