摘要
针对冲击地压监测 AE时间序列的特点 ,建立了由伸缩和平移因子决定的小波基函数代替 Sigmoid等传递函数的小波神经网络预测模型 ,避免了传统神经网络需要人为干预网络结构参数的不足。实例分析表明 ,该模型拟和预测精度高 ,具有重要的应用价值。
Based on the features of acoustic emission time series monitored for rock burst, a wavelet neural network model is put forward. In the model, the transferring function such as Sigmoid function is replaced by wavelet radix determined by flex factor and parallel factor. The model can avoid the shortcoming of changing the structure parameters artificially. It is shown by an example that the model can predict rock burst accurately, thus the model is very useful.
出处
《岩石力学与工程学报》
EI
CAS
CSCD
北大核心
2000年第z1期1034-1036,共3页
Chinese Journal of Rock Mechanics and Engineering
基金
国家自然科学基金!(5 980 40 0 5 )
山东省自然科学基金
关键词
小波神经网络
冲击地压
预测
wavelet neural network, rock burst, predictih