基于BP神经网络的桂林生态城市建设需水量预测

根据桂林市经济社会历年统计的主要指标数据,运用SPSS社会科学统计软件分析并选取出桂林市辖区生态城市建设需水量的显著性影响因子,采用改进的归一化进行非线性规格化数据处理,基于Matlab建立BP神经网络模型,预测桂林市辖区生态城市建设需水量,结果表明,预测结果与原始数据的平均相对误差为1.19%,最大为2.08%,最小为0.28%。该模型具有较高的预测精度和良好的泛化能力,BP神经网络与SPSS软件优化组合模型,可用于需水量预测。

Water demand forecast of eco-city construction in Guilin City based on BP neural network

According to statistics of the major economic and social indicators in Guilin City over recent years, significant factors of water demand for eco-city construction in Guilin City were selected using the SPSS software, and an improved normalization method was used for nonlinear normalization data processing. The MATLAB-based BP neural network model was established to forecast the water demand of eco-city construction in Guilin City. The forecasted results, compared with the original data, had an average relative error of. 1.19%, a maximum error of 2.08%, and a minimum error of 0.28 %. The model is shown to have high accuracy and good generalization ability, and the BP neural network model, combined with the SPSS software optimization, is applicable to water demand forecasting.