Improved learning algorithm for LMBP based on conjugate gradient methods

doi:10.3778/j.issn.1002-8331.2008.21.029

Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (21): 106-108.DOI: 10.3778/j.issn.1002-8331.2008.21.029

• 机器学习 • Previous Articles Next Articles

Improved learning algorithm for LMBP based on conjugate gradient methods

LI Ye-li¹,FENG Chao²,LU Li-kun¹

1.School of Information & Machine Electrical Engineering，Beijing Institute of Graphic Communication，Beijing 102600，China
2.School of Electrical Engineering and Automation，Tianjin University，Tianjin 300072，China

Received:2008-04-30 Revised:2008-05-26 Online:2008-07-21 Published:2008-07-21
Contact: LI Ye-li

一种基于共轭梯度的LMBP改进学习算法

李业丽¹,冯超²,陆利坤¹

1.北京印刷学院信息与机电工程学院，北京 102600
2.天津大学电气与自动化工程学院，天津 300072

通讯作者: 李业丽

Abstract

Abstract: The convergence of Levenberg-Marquardt BP neural network is slow.The reason is that computation costs too much in process of inversing matrix J^TJ+µI for LMBP algorithm.Based on unconstrained optimization theory，an improved LMBP learning algorithm is presented by using conjugate gradient methods.It uses the way to avoid the time-consuming calculation of the invers matrix，to decrease the calculation amount and to quicken the convergence.The validity of the algorithm is proved by the simulation in Matlab.

Key words: BP algorithm, Levenberg-Marquardt, optimization theory, conjugate gradient methods

摘要： 对神经网络中的LMBP（Levenberg-Marquardt BP）算法的收敛速度慢进行分析，针对矩阵J^TJ+µI求逆过程运算量过大而造成收敛速度慢的缺陷，根据无约束优化理论，提出一种基于共轭梯度方法的改进LMBP网络学习算法，利用求解大规模线性方程组的共轭梯度方法，避免了烦琐的求逆过程，降低了计算复杂度，加快了网络的收敛速度，通过Matlab仿真，比较了算法的收敛速度，证明了方法的有效性。

关键词: BP算法, Leverberg-Marquardt, 优化理论, 共轭梯度

LI Ye-li¹,FENG Chao²,LU Li-kun¹. Improved learning algorithm for LMBP based on conjugate gradient methods[J]. Computer Engineering and Applications, 2008, 44(21): 106-108.

李业丽¹,冯超²,陆利坤¹. 一种基于共轭梯度的LMBP改进学习算法[J]. 计算机工程与应用, 2008, 44(21): 106-108.

[1]	LI Moyu, JIAO Ruili, ZHU Yunlong, WU Shiyu. GNSS-R SAR Ground Target Modeling and Imaging [J]. Computer Engineering and Applications, 2019, 55(23): 265-270.
[2]	JIANG Guoqing1, PAN Yong2, HU Feiyue1. Research on logistics distribution route based on genetic algorithm and ant colony optimization algorithm [J]. Computer Engineering and Applications, 2015, 51(2): 255-258.
[3]	FENG Youqian1, YU Fuping2, GAO Dahua1, AN Qinli1. Research on radar shift-frequency jamming suppression based on sparse decomposition [J]. Computer Engineering and Applications, 2012, 48(6): 129-130.
[4]	SUN Juanjuan1, YU Jiguo1, LIU Xiangtao2. Similarity-computing model for P2P file-sharing networks [J]. Computer Engineering and Applications, 2012, 48(4): 111-114.
[5]	JING Junfeng1，2, ZHANG Huanhuan1, LI Pengfei1, WANG Jing1. Fabric defect classification based on local binary patterns and Tamura texture feature method [J]. Computer Engineering and Applications, 2012, 48(23): 155-160.
[6]	ZOU Hai, CHU Weicui, XU Jun, ZHANG Juan, WANG Ziqi. Application of two-weight neural networks in image segmentation [J]. Computer Engineering and Applications, 2012, 48(13): 163-166.
[7]	WANG Mingsheng1，2，LV Xikui2. Comprehensive evaluation model of engineering geology adopted improved Levenberg- Marquardt neural network [J]. Computer Engineering and Applications, 2011, 47(36): 234-237.
[8]	HU Ting1，WANG Yong1，TAO Xiaoling2. Method of network traffic classification using improved LM algorithm [J]. Computer Engineering and Applications, 2011, 47(28): 65-68.
[9]	WANG Yan-xia，DENG Wei . New text categorization model based on cascade neural network and SVD [J]. Computer Engineering and Applications, 2010, 46(26): 102-104.
[10]	XIE Wen-lan，SHI Yue-xiang，XIAO Ping. Classification of natural image based on BP neural network [J]. Computer Engineering and Applications, 2010, 46(2): 163-166.
[11]	YANG Hong^1，2，LI Ya-an¹，LI Guo-hui^1，2. New method of improved extended Kalman filter [J]. Computer Engineering and Applications, 2010, 46(19): 18-20.
[12]	YANG Hong^1，2，LI Ya-an¹，LI Guo-hui^1，2，YUAN Run-ping³. Method of improving central difference Kalman filter [J]. Computer Engineering and Applications, 2010, 46(19): 132-134.
[13]	LI Guo-hui^1，2，LI Ya-an¹，YANG Hong^1，2. Iterated central difference Kalman particle filter [J]. Computer Engineering and Applications, 2010, 46(15): 7-9.
[14]	YANG Wan-an，ZENG An-ping. New method of solving complicated nonlinear equation group [J]. Computer Engineering and Applications, 2009, 45(28): 41-42.
[15]	ZHANG Hong-yan,GENG Zheng. Novel interpretation for Levenberg-Marquardt algorithm [J]. Computer Engineering and Applications, 2009, 45(19): 5-8.

Improved learning algorithm for LMBP based on conjugate gradient methods

一种基于共轭梯度的LMBP改进学习算法

PDF

Knowledge

Abstract

Cite this article

share this article

References

Related Articles 15

Recommended Articles

Metrics