Study of Improved BP Algorithm based on Gradient Descent and Numerical Optimization

Qiuhong Sun, Weihong Bi, Xinhang Xu

2015

Abstract

Studied limitations exist in BP model, and discussed the proposed improved algorithm based on BP neural network. Respectively, from the third of the aspects discussed based on improved gradient descent algorithm and improved algorithm based on numerical optimization. Research results showed that the comprehensive method is better than the standard BP algorithm in terms of the number of iterations, the training time and the mean square error and the like, of additional momentum and adaptive outstanding performance parameter method. Researches showed that the Marquardt-Levenberg algorithm neural network convergence fastest training times at least.

References

  1. Li chungui etc. Traffic Flow forcasting Algorithm using Simulated Annealing Genetic BP Network. Proceedings of 2010 International Conference on Measuring Technology and Mechatronics Automation 2010.
  2. Shan Li, Haibing Chen, JunXian Yun etc. Optimization for Cyclosporine Blood Concentration Prediction Based on Genetic Algorithm-BP Neutal Network. Proceedings of Second International Conference on Genetic and Evolutionary Computing. 2008.
  3. Yu-min Chiang, Huei-min Chiang, Shang-yi Lin. The Application of Ant Colony Optimization for Gene Selection in Microarray-based Cancer Classification. Proceedings of Seventh International Conference on Machine Learing and Cybernetics. 2008.
  4. Li Wang, Dong-qing Wang, Ning Ding. Research on BP Neural Network Optimal Method Based on Improved Ant Colony Algorithm. IEEE Computer Society Proceedings of 2010 Second International Conference on Computer Engineering and Applications. 2010.
  5. Yan Zhao, Zhongjun Xiao, Jiayu Kang. Optimization Design Based on Improved Ant Colony Algorithm for PID Parameters of BP Neural Network. Proceedings of 2010 2nd International Asia Conference on Informatics in control, Automation and Robotics. 2010.
  6. Yuxiang Shao, Qing Chen. Application Ant Colony Neural Network in Lithology Recognition And Prediction:Evidence from China. Proceedings of 2008 IEEE Pacific-Asia Workshop on Computational Intelligence and Industrial Applications. 2008.
  7. Wu Yuguo, Song Chongzhi. Design Optimization Based on Neural Networks and Ant Colony Optimization. Proceedings of 2007 Second IEEE Conference on Industrial Electronics and Applications. 2007.
  8. Yu-min Chiang, Huei-min Chiang, Shang-yi Lin. The Application of Ant Colony Optimization for Gene Selection in Microarray-based Cancer Classification. Proceedings of Seventh International Conference on Machine Learing and Cybernetics. 2008.
Download


Paper Citation


in Harvard Style

Sun Q., Bi W. and Xu X. (2015). Study of Improved BP Algorithm based on Gradient Descent and Numerical Optimization . In Proceedings of the Information Science and Management Engineering III - Volume 1: ISME, ISBN 978-989-758-163-2, pages 452-456. DOI: 10.5220/0006028404520456


in Bibtex Style

@conference{isme15,
author={Qiuhong Sun and Weihong Bi and Xinhang Xu},
title={Study of Improved BP Algorithm based on Gradient Descent and Numerical Optimization},
booktitle={Proceedings of the Information Science and Management Engineering III - Volume 1: ISME,},
year={2015},
pages={452-456},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0006028404520456},
isbn={978-989-758-163-2},
}


in EndNote Style

TY - CONF
JO - Proceedings of the Information Science and Management Engineering III - Volume 1: ISME,
TI - Study of Improved BP Algorithm based on Gradient Descent and Numerical Optimization
SN - 978-989-758-163-2
AU - Sun Q.
AU - Bi W.
AU - Xu X.
PY - 2015
SP - 452
EP - 456
DO - 10.5220/0006028404520456