Comparative analysis of feed forward and radial basis function neural networks for the reconstruction of noisy curves

Neural networks are considered to be an important tool for interpolation and curve fitting problems. Two important neural networks- the multi-layer feed forward network and the radial basis function network (RBF) are considered for fitting of noisy curves. Comparison between the two networks is drawn on the basis of noise in the data. Performance is shown for varying levels of noise and thus the conclusions are drawn on the suitability of the two networks for the problem of reconstruction of noisy curves.