An algorithm for fast convergence in training neural networks

Authors:
  • BWB.M. Wilamowski
    B.M. Wilamowski
  • SIS. Iplikci
    S. Iplikci
  • OKO. Kaynak
    O. Kaynak
  • MEM.O. Efe
    M.O. Efe

Abstract
In this work, two modifications on Levenberg-Marquardt algorithm for feedforward neural networks are studied. One modification is made on performance index, while the other one is on calculating gradient information. The modified algorithm gives a better convergence rate compared to the standard Levenberg-Marquard (LM) method and is less computationally intensive and requires less memory. The performance of the algorithm has been checked on several example problems.

This publication has 12 references indexed in Scilit: