基于BP神经网络的太湖富营养化时空变化预测

Spatiotemporal Changes Prediction of Taihu Lake Based on BP Neural Network

建立拓扑结构为5-9-1的3层BP神经网络模型,预测太湖的富营养化水平,为水环境监测平台提供决策支持。该模型采用动量梯度下降法对太湖的湖体的富营养化水平进行预测并评价。验证结果表明,训练后的网络得出的预测结果与实际值很接近。因此,采用BP(Back Propagation)神经网络对太湖湖体富营养化水平进行预测并评价是一种有效的方法。

BP(Back Propagation) artificial neural network has excellent nonlinear approximation ability and can be used for prediction.In this paper,a BP neural network model with the topology of 5-9-1 is established for prediction and evaluation.Test results showed that this model is an effective way in predicting the eutrophication level of Taihu Lake,and the error between network output value and actual value is very little,so this model can be used in routine work for prediction and give necessary support to the water environment platform.