摘要
在混沌理论和相空间重构技术的基础上,提出了一种基于小生境自适应差分进化小波神经网络(NADE-WNN)的混沌背景下弱信号检测方法。该方法采用小生境自适应差分进化算法同时优化小波神经网络的结构和参数,简化网络结构,提高网络的学习精度和收敛速度。实验结果表明,与传统的RBF神经网络和小波神经网络预测混沌时间序列的性能相比,该算法优化的小波神经网络具有更高的预测精度和收敛速度,能够较好地检测出混沌背景下的弱信号。
A novel method of weak signal detection in chaos condition of niche adaptive differential evolution wavelet neural network (NADE-WNN) model was presented based on chaos theory and phase-space reconstruction technology. The structures and parameters of wave- let neural network were optimized by NADE algorithm at the same time in the model, the network structure was simplified and the learning precision as well as convergence rate were improved. Comparing with traditional RBF neural network and wavelet neural network, the experimental results have shown that the NADE-optimized wavelet neural network has the performance of higher prediction accuracy and convergence rate in chaotic time series prediction, and can better detect the weak signal in the background of chaos.
出处
《计算机应用与软件》
CSCD
2010年第3期29-31,39,共4页
Computer Applications and Software
基金
国家自然科学基金项目(50677014
60876022)
湖南省自然科学基金项目(06JJ2024)
湖南省教育厅科研项目(05C188)
关键词
小生境
自适应
差分进化算法
小波神经网络
弱信号
检测
Niche Adaptive Differential evolution algorithm Wavelet neural network Weak signal Detection