Non Monotone Backtracking Inexact BFGS Method for Regression Analysis

Abstract

In this article, we propose a non monotone backtracking line search technique combining with BFGS (Broyden et al., 1970) method for solving regression problems. The convergent results of the given method are established under mild conditions. Numerical results show that the proposed algorithm is interesting.

Keywords:

• BFGS method
• Fitting method
• Global convergence
• Non monotone technique
• Regression analysis

Mathematics Subject Classification:

• 65K05

Acknowledgment

This work is supported by the Program for Excellent Talents in Guangxi Higher Education research Project (Grant No. 201010LX288), Guangxi NSF (Grant No. 2012GXNSFAA053002), and China NSF (Grant Nos. 11261006, 11161003 and 71001015).