Dynamic reconstruction of sea clutter using regularized REP networks

We demonstrate the dynamic reconstruction of sea clutter time series using a regularized radial basis function (RBF) network. The dynamic invariants, namely, correlation dimension, Lyapunov exponents and the Kaplan-Yorke dimension of the actual and the reconstructed time series are compared to confirm the goodness of fit of the RBF model to the actual clutter data. A detailed statistical analysis of the horizon of predictability is presented to show the model´s ability to capture the underlying dynamics of sea clutter. These convincing results show that the RBF model is capable of approximating the dynamics of sea clutter process in a convincing manner.