[1]PU Xingcheng,LIN Yanqin.Hybrid pruning algorithm for the neural network based on significance analysis[J].CAAI Transactions on Intelligent Systems,2014,9(6):690-697.[doi:10.3969/j.issn.1673-4785.201309062]
Copy

Hybrid pruning algorithm for the neural network based on significance analysis

CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume: 9 Number of periods: 2014 6 Page number: 690-697 Column: 学术论文—机器学习 Public date: 2014-12-25
Title:
Hybrid pruning algorithm for the neural network based on significance analysis
Author(s):
PU Xingcheng1; 2;  LIN Yanqin1
1. Department of Computer Science, Chongqing University of Post & Telecommunications, Chongqing 400065, China;
2. Department of Mathematics & Physics, Chongqing University of Post & Telecommunications, Chongqing 400065, China
Keywords:
significance analysis; neural network; cooperative co-evolutionary genetic algorithms; pruning algorithm; stock market
CLC:
TP24
DOI:
10.3969/j.issn.1673-4785.201309062
Abstract:
This paper puts forward a kind of hybrid pruning algorithm for considering the problem of neural network structure design. Firstly, the algorithm uses the different advantages of cooperative co-evolutionary genetic algorithm and back propagation algorithm to optimize the structure and weights of neural networks. Secondly, by calculating the significance of the hidden layer neurons, it prunes the network that is not significant, further simplifying the structure of the network without reducing the generalization ability of the model. Finally, the proposed hybrid pruning algorithm is used to forecast the stock market. The simulations showed that the improved algorithm has better generalization ability and higher fitting precision than other optimization algorithms.
References:
[1] YONG Mingjing, HUA Yingdong. Study on characteristic of fractional master-slave neural network[C]//Fifth International Symposium on Computational Intelligence and Design. Hangzhou, China, 2012: 498-501.
[2] 乔俊飞, 张颖. 一种多层前馈神经网络的快速修剪算法[J]. 智能系统学报, 2008, 3(2): 622-627. QIAO Junfei, ZHANG Ying. Fast unit pruning algorithm for feed forward neural network design [J]. CAAI Transactions on Intelligent Systems, 2008, 3(2): 622-627.
[3] MOOD Y J. Prediction risk and architecture selection for neural networks[C]//Statistics to Neural Networks: Theory and Pattern Recognition Applications, NATO ASI Series F. New York, 1994: 288-290.
[4] LIU Yong. Create stable neural networks by cross-validation[C]//International Joint Conference on IJCNN’06 Neural Networks. Vancouver, Canada, 2006: 3925-3928.
[5] DU Juan, ER Mengjoo. A fast pruning algorithm for an efficient adaptive fuzzy neural network[C]//Eighth IEEE International Conference on Control and Automation. Xiamen, China, 2010: 1030-1035.
[6] 乔俊飞, 韩红桂. 前馈神经网络分析与设计[M]. 北京:科学出版社, 2013: 107-158.
[7] EUGENE S, MARIA S. A self-configuring genetic programming algorithm with modified uniform crossover[C]//WCCI 2012 IEEE World Congress on Computational Intelligence. Brisbane, Australia, 2012: 1984-1990.
[8] 吴鹏. 基于语法引导的遗传编程在神经树中的应用[D]. 济南: 济南大学, 2007: 69-102. WU Peng. The application of genetic programming in the neural tree based on grammar guide[D]. Jinan: University of Jinan, 2007: 69-102.
[9] 周明, 孙树栋. 遗传算法原理及应用[M]. 北京:国防工业出版社, 1999: 1-26.
[10] 巩敦卫, 孙晓燕. 基于合作式协同进化算法的神经网络优化[J]. 中国矿业大学学报, 2006, 35(1): 114-119.GONG Dunwei, SUN Xiaoyan. Optimization of neural network based on cooperative co-evolutionary algorithm [J]. Journal of China University of Mining & Technology, 2006, 35(1): 114-119.
[11] POTTER M A. The design and analysis of a computational model of cooperative co-evolution [D]. Fairfax: George Mason University, 1997: 21-66.
[12] HETCHT N R. Theory of the back propagation neural network[C]//Proceedings of the International Joint Conference on Neural Networks. New York: IEEE Press, 1989: 593-611.
[13] BLUM E K, LI L K. Approximation theory and feed forward networks[J]. Neural Net, 1991,2(3): 511-515.
Similar References:

Memo

-

Last Update: 2015-06-16

Copyright © CAAI Transactions on Intelligent Systems