| 基于Sanger算法的小波网络结构优化方法 |
| |
| 引用本文： | 刘守生,于盛林,丁勇. 基于Sanger算法的小波网络结构优化方法[J]. 计算机工程与设计, 2004, 25(9): 1438-1440,1456 |
| |
| 作者姓名： | 刘守生 于盛林 丁勇 |
| |
| 作者单位： | 南京航空航天大学,自动化学院,江苏,南京,210016;解放军理工大学,理学院,江苏,南京,210007;南京航空航天大学,自动化学院,江苏,南京,210016 |
| |
| 基金项目： | 国家自然科学基金(60234010) |
| |
| 摘 要： | 为了解决离散小波神经网络(DWNN)节点过多、鲁棒性差的问题，基于主成份分析(PCA)的思想提出了一种规模小、抗干性强的广义小波神经网络(EWNN)，并利用Sanger算法对其结构进行了优化。该算法在引出了消冗变换后，可提取出多个主成份。仿真结果表明了EWNN的非线性逼近能力及稳定性都明显优于DWNN。
|
| 关 键 词： | 广义小波神经网络 小波框架 主成份分析 Sanger算法 |
| 文章编号： | 1000-7024(2004)09-1438-03 |
Structure optimal method of wavelet network based on sanger algorithm |
| |
| LIU Shou-sheng,YU Sheng-lin,DING Yong College of Automation,Nanjing University of Aeronautics and Astronautics,Nanjing ,China College of Science of PLA University of Science and Technology,Nanjing ,China. Structure optimal method of wavelet network based on sanger algorithm[J]. Computer Engineering and Design, 2004, 25(9): 1438-1440,1456 |
| |
| Authors: | LIU Shou-sheng YU Sheng-lin DING Yong College of Automation Nanjing University of Aeronautics Astronautics Nanjing China College of Science of PLA University of Science Technology Nanjing China |
| |
| Affiliation: | LIU Shou-sheng,YU Sheng-lin,DING Yong College of Automation,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China College of Science of PLA University of Science and Technology,Nanjing 210007,China |
| |
| Abstract: | In order to resolve the problems of discrete wavelet neural network (DWNN), such as much nods, lack of robustness, anextended wavelet neural network (EWNN), which has small scale of nods and high quality of anti-interference, is designed on the basisof theory of principal component analysis (PCA), and its structure is optimized through sanger algorithm. Many principal componentscan be obtained after deflation. The simulation results show that EWNN is saperior in nonlinear approximation and stability to the DWNN. |
| |
| Keywords: | extended wavelet neural network wavelet frame PCA sanger algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
