Performance comparison between back propagation, RPE and MRPE algorithms for training MLP networks

• Autores: Mohd Yusoff Mashor
• Localización: International journal of the computer, the internet and management, ISSN 0858-7027, Vol. 8, Nº. 3, 2000, págs. 65-78
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • This paper presents the performance comparison between back propagation, recursive prediction error (RPE) and modified recursive prediction error (MRPE) algorithms for training multilayered perceptron networks. Back propagation is a steepest descent type algorithm that normally has slow convergence rate and the search for the global minimum often becomes trapped at poor local minima. RPE and MRPE are based on Gaussian-Newton type algorithm that generally provides better performance. The current study investigates the performance of three algorithms to train MLP networks. Two real data sets were used for the comparison. Its was found that the RPE and MRPE are much better than the back propagation algorithm.