利用多分辨分析方法,结合小波分析和神经网络思想构建一种新型的神经网络模型———小波神经网络,解决了传 统神经网络中隐层节点数难以确定的问题。通过对股票的预测,说明该方法能有效地提高预测精度, 避免了人工神经网 络模型的固有缺陷。 -Using multi-resolution analysis method, combined with wavelet analysis and neural network ideological construct a new neural net

基于小波变换的图像纹理特征提取方法及其应用-A Method of Image Texture Feature Analysis Based on Wavelet Decomposition and its Application

A fractional Fourier transform (FrFT) based estimation method is introduced in this paper to analyze the long range dependence (LRD) in time series. The degree of LRD can be characterized by the Hurst parameter. The FrFTbased estimation of Hu

通过小波分析和独立分量分析来去除脑电信号中的运动伪迹，有很好的参考价值。-Wavelet analysis and independent component analysis to remove EEG motion artifact, a good reference value.(Robust removal of short-duration artifacts in long neonatal EEG recordings using wavelet-enhanced ICA and