Weak Signal Detection in Noisy Chaotic Time Series Using ORBFNN

This paper considers the problem of detection of weak signal detection in noisy chaotic time series using an optimal radial basis function neural network (ORBFNN). Based on chaotic dynamic mechanism, using ORBFNN to establish the forecast model of chaotic time series. When noise exists, to determine the structure of an optimal RBF predictor, we propose a new technique called the cross-validated subspace method to estimate the optimum number of hidden units. Which is used to identify a suitable number of hidden units by detecting the dimension of the subspace spanned by the signal eigenvectors, the cross validation method is applied to prevent the problem of overfitting. The results of theoretical analysis and simulation indicate the effectiveness of the ORBFNN predictor. The infection degree of noise is evaluated in quantity in the end. Results show that the proposed ORBFNN predictor can provide a further improvement in signal detection performance.