河流水质预测的灰色神经网络联合模型

Prediction of river water quality using the united gray neural network model

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,运用响应成分模型将水质浓度分解为具有确定性的趋势项和具有不确定性的随机项,建立灰色神经网络联合模型对水质的趋势项和随机项进行预测。将该模型应用于淮河河流水质预测,结果表明,该模型拟合误差小,预测精度高。

Aiming at the defect that the gray method can only predict the tendency approximately, the coneentration of contaminant was decomposed into trend item and random item based on the responding eomposition model. The trend items and random items of concentration of contaminant time series were predicted by united gray and neural network model. As an example, the united gray and neural network model was applied to the prediction of the river water quality. The results showed this model was the valid and feasible and high precision.