An Intelligent Method Optimizing BP Neural Network Model

Abstract:

How to build the model represents the complex relationship of data and how to optimize the model has been the core issue of data mining research. BP neural network was able to characterize the nonlinear data relationships by training. But BP neural network is easy to fall into local minimum, and its hidden nodes, the connection weights and thresholds are not easy to determine. To overcome the shortcomings of the BP neural network model, this paper presents an intelligent method based on genetic algorithm optimizing the BP neural network according to the error minimization principle. Experimental results with function approximation and remote sensing image classification indicate that the optimized model can be an effective way to improve forecasting accuracy.

• Periodical: Advanced Materials Research (Volumes 605-607)
• Pages: 2470-2474
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.605-607.2470
• Online since: December 2012
• Authors: Hai Hua Xing, Hong Yan Lin
• Keywords: Artificial Neural Network (ANN), Genetic Algorithm (GA), Model Optimized

References

[1] Liaoqin, Hao Zhifeng, Chen Zhihong.Data mining and mathematical modeling[M]. National Defence Industry Press,2010.

[2] Shiyan, Han Liqun, Lian Xiaoqin. Neural Network Design and Case Analysis [M].Beuing University of Posts and Telecommunications Press,2009.

[3] Tang Wanmei. The Study of the Optimal Structure of BP Neural Network [J]. Systems Engineering-theory & Practice, 2005(10)95:100．

[4] He Yong, Xiang Liguo. Study and application of BP neural network model based on fuzzy clustering [ J] . Systems Engineering-Theory & Practice, 2004, 24( 2) : 79- 82.

[5] ZOU Hao-fei,XIA Guo-ping,YANG Fang-ting. Neural network forecasting model using multi-stage optimization approach based on GMDH and genetic algorithm[J]. Management Sciences in China,2006(5)

[6] Li Xiaofeng, Xu Jiuping,Wang Yinqing , et al. The establishmen t of sel- f adapting algorithm of BP neural network and its application[ J] . Systems Engineering - Theory & Practice, 2004, 24( 5) : 1- 8.

[7] Wang Xiaoping, Cao Liming. Genetic algorithm - theory, application and software implementation[M]. Xi'an Jiaotong University Press, 2002.

[8] Lam H K, et al. Tuning of the Structure and Parameters of Neural Network Using an Improved Genetic Algorithm[ C] . Industrial Electronics Society, Denver, CO, USA: IECON) 01, 2001.

DOI: 10.1109/iecon.2001.976448

[9] Landgrebe D A. Signal theory methods in multispectral remote sensing[M]. Hoboken, NJ: Wiley Interscience, 2003.