摘要
针对水质参数样本数据少且非线性的特点,建立了新陈代谢无偏GM(1,1)与BP神经网络的组合预测模型,将通过新陈代谢无偏GM(1,1)模型得到的数据集作为BP神经网络的输入,原始序列作为神经网络的期望输出,训练得到最佳BP神经网络。将该组合模型应用于乐山岷江大桥断面溶解氧浓度的预测,结果表明,相对误差均在3%以下,与传统灰色神经网络水质预测模型相比,该模型具有实时性及预测精度更高的优点。
In view of small sample of water quality data with nonlinear feature,combined information renewal unbiased GM（1,1） and BP neural network forecast model is proposed.Using the data set preprocessed by information renewal unbiased GM（1,1） as the input of the BP neural network,and the original data as the expected output,the BP neural network is trained to get the optimal structure.The combined model is applied to forecast the DO concentration of section of Minjiang Bridge.Simulation results show that compared with the traditional grey neural network model,the proposed forecast model is of real-time capability and superior in forecasting precision with relative error below 3%.
出处
《水电能源科学》
北大核心
2012年第2期35-37,共3页
Water Resources and Power
