详细信息
第二类Chebyshev前向神经网络权值直接确定及结构自适应确定 被引量:7
Weights-direct-determination and structure-adaptive determination of feed-forward neural network activated with the 2nd-class Chebyshev orthogonal polynomials
文献类型:期刊文献
中文题名:第二类Chebyshev前向神经网络权值直接确定及结构自适应确定
英文题名:Weights-direct-determination and structure-adaptive determination of feed-forward neural network activated with the 2nd-class Chebyshev orthogonal polynomials
作者:肖秀春[1,2];张雨浓[2];姜孝华[2];邹阿金[1,2]
机构:[1]广东海洋大学信息学院,广东湛江524088;[2]中山大学信息科学与技术学院,广州510275
年份:2009
卷号:35
期号:1
起止页码:80
中文期刊名:大连海事大学学报
外文期刊名:Journal of Dalian Maritime University
收录:CSTPCD、、Scopus、北大核心2008、CSCD_E2011_2012、北大核心、CSCD
基金:国家自然科学基金资助项目(60643004;60775050);中山大学科研启动费、后备重点课题资助项目
语种:中文
中文关键词:神经网络;正交多项式;权值直接确定;网络结构;自适应
外文关键词:neural network; orthogonal polynomials; weights direct determination ; network structure; adaptation
中文摘要:为克服BP神经网络模型及其学习算法中的固有缺陷,构造了第二类Chebyshev前向神经网络模型,提出该神经网络模型权值直接确定法和结构自适应确定法.理论分析及仿真实验均表明,该系统弥补了BP神经网络的某些固有缺陷.相比同构型BP神经网络,其计算速度和工作精度均有大幅提高.
外文摘要:To remedy the weaknesses of back propagation(BP) neural network model and its learning algorithm, a weights-di- rect-determination and structure-adaptive-determination method of feed-forward neural network activated with the 2nd -class Chebyshev orthogonal polynomials was developed. Theoretical analysis and simulation results both show that the proposed sys- tem can remedy the weaknesses of BP neural network model and its learning algorithm, and the calculation speed and working precision improve a lot comparing with the same structured BP neural network.
参考文献:
正在载入数据...