有机灰色神经网络模型在河流水质预测中的应用

Application of Organic Gray Neural Network Model to River Water Quality Prediction

针对灰色预测对波动较强的序列只能预测大致变化的缺陷,在分析河流水质动态变化的基础上,结合灰色理论中的GM(1,1),无偏GM(1,1)和RBF神经网络的特点,提出有机灰色神经网络预测模型,将灰色模型得到的数值作为神经网络的输入,原始数据作为神经网络的输出,训练得到最佳神经网络结构.以某地区河流水质为例,根据其变化规律,应用有机灰色神经网络模型进行预测,结果表明,该模型拟合误差小,预测精度高.

In view of the defect that the gray method can only predict the tendency approximately, a new organic gray neural network model is proposed by the advantages of GM（1,1）, gray residual difference identifi- cation and RBF neural network, based on the analysis of the river water quality. The three groups data got from the gray model is used as the input of the neural network and the origin data are used as the output of neural network. The neural network is trained to get the optimal structure of neural network. According to the dynamic law of a certain river water quality in some region, the water quality is predicted by using organic gray neural network model. The results show that the model has highly fitting and predicting precision advantages than other model.