CAJ | 학술논문

构建小波与人工神经网络组合时序模型，该模型利用 morlet 小波基函数取代人工神经网络的激发函数，通过平移因子和伸缩因子确定小波基函数，采用误差逆向传播算法训练网络和预测。利用此模型对赣江河段上外洲水文站月径流变化趋势进行预测，并将其计算结果与 BP 神经网络进行比较，结果表明利用小波神经网络进行时间序列预测效果较理想。
구건소파여인공신경망락조합시서모형，해모형이용 morlet 소파기함수취대인공신경망락적격발함수，통과평이인자화신축인자학정소파기함수，채용오차역향전파산법훈련망락화예측。이용차모형대공강하단상외주수문참월경류변화추세진행예측，병장기계산결과여 BP 신경망락진행비교，결과표명이용소파신경망락진행시간서렬예측효과교이상。
In this paper,wavelet analysis and artificial neural networks are combined into a timing model. The model uses morlet wavelet function to replace artificial neural network excitation function and back-ward error propagation algorithm to train the network and prediction. It is used to predict the annual runoff characteristics of Waizhou hydrological station in Ganjiang River. The result shows that compared with BP artificial neural networks,artificial neural networks prove to be a more effective way.