Regress ion based weight generation algorithm in neural network for estimation of frequencies of vibrating plates

This paper introduces a new algorithm for the neural network training in order to have efficient learning and training of the network. The proposed methodology has been used to estimate the vibration characteristics of plate structures. The neural network has been trained using the weights generated by the coefficients of regression polynomials of particular degrees. The benefit of using the proposed approach as compared to exiting approaches viz. the neural network toolbox in MATLAB has also been investigated. Efficacy and powerfulness of the new algorithm has been tested for the mentioned example problem and the results when tested with the new data have been found to be in good agreement. (c) 2005 Elsevier B.V. All rights reserved.