A Neural Network Training Algorithm Based on Collinear Scaling Quasi-Newton Method

Author

Ye, Shijie ; Li, Jianliang ; Xu, Jun

Author_Institution

Sch. of Sci., Nanjing Univ. of Sci. & Technol., Nanjing, China

fYear

2011

fDate

25-27 April 2011

Firstpage

139

Lastpage

141

Abstract

Based on collinear scaling and local quadratic approximation, quasi-Newton methods have improved for function value is not fully used in the Hessian matrix. As collinear scaling factor in paper may appear singular, this paper, a new collinear scaling factor is studied. Using local quadratic approximation, an improved collinear scaling algorithm to strengthen the stability is presented, and the global convergence of the algorithm is proved. In addition, numerical results of training neural network with the improved collinear scaling algorithm shown the efficiency of this algorithm is much better than traditional ones.

Keywords

Newton method; approximation theory; convergence; learning (artificial intelligence); Hessian matrix; collinear scaling algorithm; collinear scaling factor; function value; global convergence; local quadratic approximation; neural network training; quasiNewton method; stability; Algorithm design and analysis; Approximation algorithms; Approximation methods; Artificial neural networks; Convergence; Optimization; Training; Neural network training; collinear scaling; local quadratic approximation; numberical analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Information and Computing (ICIC), 2011 Fourth International Conference on

Conference_Location

Phuket Island

Print_ISBN

978-1-61284-688-0

Type

conf

DOI

10.1109/ICIC.2011.22

Filename

5954523