计算机与数字工程
計算機與數字工程
계산궤여수자공정
Computer and Digital Engineering
2015年
9期
1557-1560
,共4页
小波神经网络%训练算法%人口预测%自适应
小波神經網絡%訓練算法%人口預測%自適應
소파신경망락%훈련산법%인구예측%자괄응
wavelet neural network%training algorithm%population projections%adaptive
小波神经网络是小波分析理论和人工神经网络两者结合的产物。为了解决 BP 算法易陷入极小值,收敛速度慢的问题。论文采用小波神经网络进行训练同时融合了带动量的 BP 算法,自适应调整小波的伸缩参数、平移参数和连接权值,并在其选定的范畴内尽可能多地提取信号特征。最后,论文以人口预测为例验证了自适应小波神经网络算法有效性。
小波神經網絡是小波分析理論和人工神經網絡兩者結閤的產物。為瞭解決 BP 算法易陷入極小值,收斂速度慢的問題。論文採用小波神經網絡進行訓練同時融閤瞭帶動量的 BP 算法,自適應調整小波的伸縮參數、平移參數和連接權值,併在其選定的範疇內儘可能多地提取信號特徵。最後,論文以人口預測為例驗證瞭自適應小波神經網絡算法有效性。
소파신경망락시소파분석이론화인공신경망락량자결합적산물。위료해결 BP 산법역함입겁소치,수렴속도만적문제。논문채용소파신경망락진행훈련동시융합료대동량적 BP 산법,자괄응조정소파적신축삼수、평이삼수화련접권치,병재기선정적범주내진가능다지제취신호특정。최후,논문이인구예측위례험증료자괄응소파신경망락산법유효성。
Wavelet neural network is the product combined with wavelet analysis theory and artificial neural network . In order to solve the problem of BP algorithm easily falling into the minimum ,slow convergence .In this paper ,wavelet neu‐ral network is trained to drive while incorporating the amount of BP algorithm ,adaptive wavelet stretching parameters , translation parameters and connection weights ,and extract the signal features in their selected areas as much as possible .Fi‐nally ,an example to the population forecast is made to verify the effectiveness of adaptive wavelet neural network algorithm .