CAJ | 학술논문

运用BP神经网络理论和方法，建立了河南省新乡市百泉泉水流量动态的BP神经网络模型，并与逐步回归模型和灰色模型进行了比较，结果表明：BP神经网络模型的预测值与实测值最为接近，拟合效果明显优于逐步回归模型和灰色模型的拟合效果，三者的相对误差分别为0．961％、10．734％、5．664％。根据三种模型的预测结果，利用周期分析法和补偿开采的观点评价百泉泉域裂隙岩溶地下水可开采资源为3．25 m3/s。
운용BP신경망락이론화방법，건립료하남성신향시백천천수류량동태적BP신경망락모형，병여축보회귀모형화회색모형진행료비교，결과표명：BP신경망락모형적예측치여실측치최위접근，의합효과명현우우축보회귀모형화회색모형적의합효과，삼자적상대오차분별위0．961％、10．734％、5．664％。근거삼충모형적예측결과，이용주기분석법화보상개채적관점평개백천천역렬극암용지하수가개채자원위3．25 m3/s。
This paper applied BP neural network theory and method to establish dynamic BP neural network model of spring flow of Baiquan at Xinxiang City of Henan Province and compared with the predicted results of the stepwise regression model and gray model. The results show that the predictive value of the BP model is most close to the measured value and its fitting effect is obviously better than the stepwise regression model and gray model. The relative errors of the three models are 0. 961%,10. 734% and 5. 664% respectively. According to the predicted results of the three models,the groundwater exploitation resources of fissure karst at Baiquan spring territory evaluated with the periodic analysis method and the view point of compensation mining is 3. 25 m3/s.