BP Neural Network Research Based on Three Convergence Improved LM Algorithm


Article Preview

In order to solve the practical application problem, which traditional neural network takes too long and compute complexly, on the basis of the LM algorithm, combined with mathematical optimization theory, identify the three convergence Improved LM algorithm applied to BP neural network , that improved LMBP algorithm. Simulation results show that the improved LMBP algorithm in convergence time and goodness of fit both have better results, and the algorithm is general and can be produced by obtaining national sample of various scenarios, using the algorithm to predict, to better guidance on production.



Edited by:

Yun-Hae Kim and Prasad Yarlagadda




X. C. Guo and S. H. Shang, "BP Neural Network Research Based on Three Convergence Improved LM Algorithm", Applied Mechanics and Materials, Vols. 303-306, pp. 1543-1546, 2013

Online since:

February 2013




[1] Shouren Hu. Introduction to Neural Networks[M]. Beijing: National Defense University Press, 1993.

[2] Daikui. Neural network application technology [M] Beijing: National Defense University Press, (1998).

[3] Pingfan Yan. Multilayer neural network research[J]. AAS, 1997 , 23(1 ): 129 -135.

[4] J.E. Dennis Jr.,D.M. Gay and E. Welsch.: An adaptive nonlinear least-squares algorithm-ACM Transactions on Math. software, -1981, 7: 348-383.

DOI: https://doi.org/10.1145/355958.355966

[5] M.C. Bartholomew-Biggs.: The estimation of the Hessian matrix in nonlinear least squares problems with non-zero residua-ls. Math. Prog., 1977, 12: 67-80.

DOI: https://doi.org/10.1007/bf01593770