Estimation and representation of non-linear static functions using non-orthogonal continuous wavelets

This paper presents a wavelet neural-network for learning and approximation of non-linear functions. Wavelet networks are a class of neural network that takes advantage of good localization and approximation properties of multiresolution analysis. The proposed model structure is similar to that of Radial Basis Function (RBF) structure and we have restricted to non-orthogonal continuous wavelets i.e. the first and second derivative of Gaussians as wavelet functions. Simulations show that the proposed method outperforms the other reported methods.