ABSTRACT
This chapter explores numerical results are presented to substantiate the efficacy and superior abilities of the presented Chebyshev-polynomial-of-Class-2 neuronet in terms of approximation, testing, denoisingdenoising and prediction, with the aid of the weights-and-structure-determination (WASD) algorithm which obtains the optimal number of hidden-layer neurons of Chebyshev-polynomial-of-Class-2 neuronet. Numerical studies further substantiate the efficacy and superior abilities of Chebyshev-polynomial-of-Class-2 neuronet in approximation, denoising and prediction, with the aid of the WASD algorithm which obtains the optimal number of hidden-layer neurons. In order to avoid the usually lengthy iterative-training procedure and improve the efficacy of Chebyshev-polynomial-of-Class-2 neuronet, a weights-and-structure-determination WASD algorithm is designed for such a neuronet. In view of the advantages of the weights-direct-determination subalgorithm and structure-automatic-determination subalgorithm, a WASD algorithm to obtain efficiently the optimal number of hidden-layer neurons of the two-inputtwo-input Chebyshev-polynomial-of-Class-2 neuronet. Numerical studies further substantiate the efficacy and superior performance of Chebyshev-polynomial-of-Class-2 neuronet equipped with the WASD algorithm in approximation, denoisingdenoising and prediction.
