TPPMA: New Adaptive BP Neural Network Based on PSO and PCA Algorithms

In this paper, we investigate the optimization of BP neural network. An optimized BP neural network model, adaptive BP neural network based on PSO and PCA algorithms (TPPMA) is proposed to improve the training speed and increase prediction accuracy. By introducing Momentum backpropagation and adaptive learning rate into BP Neural Network, one can reduce the possibility of local optimization. For the architecture of BP Neural Network, the initial connection weight is determined by Particle Swarm Optimization (PSO) and the number of hidden nodes is decided by three-division method. In the model, Principal Component Analysis (PCA) is used to reduce the dimension of the sample. Simulation result demonstrates that TPPMA method is efficient and can get promising results with less time compared with other results.

[1]  Yong Wang,et al.  Feature selection using tabu search with long-term memories and probabilistic neural networks , 2009, Pattern Recognit. Lett..

[2]  Jasmina Arifovic,et al.  Using genetic algorithms to select architecture of a feedforward artificial neural network , 2001 .

[3]  George-Christopher Vosniakos,et al.  Optimizing feedforward artificial neural network architecture , 2007, Eng. Appl. Artif. Intell..

[4]  Teresa Bernarda Ludermir,et al.  An Optimization Methodology for Neural Network Weights and Architectures , 2006, IEEE Transactions on Neural Networks.

[5]  Xin Yao,et al.  A constructive algorithm for training cooperative neural network ensembles , 2003, IEEE Trans. Neural Networks.

[6]  Zhang Min.  BP Neural Network with Adaptive Learning Rate , 2006 .

[7]  Tim Verdonck,et al.  Robust kernel principal component analysis and classification , 2010, Adv. Data Anal. Classif..

[8]  Xiaogang Ruan,et al.  A tabu based neural network learning algorithm , 2007, Neurocomputing.

[9]  Chenglei Yang,et al.  A BP Neural Network Prediction Model of the Urban Air Quality Based on Rough Set , 2008, 2008 Fourth International Conference on Natural Computation.

[10]  Xin Yao,et al.  A New Adaptive Merging and Growing Algorithm for Designing Artificial Neural Networks , 2009, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[11]  Hitoshi Iba,et al.  Learning polynomial feedforward neural networks by genetic programming and backpropagation , 2003, IEEE Trans. Neural Networks.

[12]  Guoqiang Peter Zhang,et al.  Neural networks for classification: a survey , 2000, IEEE Trans. Syst. Man Cybern. Part C.

[13]  Chen Qiu,et al.  The fault diagnosis expert system based on BP neural network , 2015, 2015 IEEE Advanced Information Technology, Electronic and Automation Control Conference (IAEAC).

[14]  Guang Yang,et al.  Application of the rough set theory and BP neural network model in disease diagnosis , 2010, 2010 Sixth International Conference on Natural Computation.

[15]  Gaofeng Zheng,et al.  Convergence analysis of a back-propagation algorithm with adaptive momentum , 2011, Neurocomputing.

[16]  B.H.M. Sadeghi,et al.  A BP-neural network predictor model for plastic injection molding process , 2000 .

[17]  Hui Qin Yao,et al.  Based on the Genetic Algorithm to Optimize the BP Neural Network in the Degree of Concrete Creep Prediction Model , 2014 .

[18]  Kazuyuki Murase,et al.  A new algorithm to design compact two-hidden-layer artificial neural networks , 2001, Neural Networks.

[19]  J. C. Pereira,et al.  Unsupervised component analysis: PCA, POA and ICA data exploring - connecting the dots. , 2016, Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy.

[20]  Su Yan-mang.  Fault Diagnosis in Analog Circuits Based on Neural Networks , 2007 .

[21]  Haipeng Pan,et al.  A new intelligent decoupling controller based on BP neural network PID with predictive compensation and the research on its application to an air-cushioned headbox , 2010, 2010 8th World Congress on Intelligent Control and Automation.

[22]  Zexuan Zhu,et al.  A fast pruned-extreme learning machine for classification problem , 2008, Neurocomputing.