Adaptive noise smoothing method with neural networks

Combing with gradient decent algorithm in neural networks, an adaptive chaotic noise smoothing method is proposed. In this paper, wavelet coefficients of chaotic signals including approximate and detail information are obtained with dual-lifting wavelet transform. The approximate parts are handled by singular spectrum analysis in order to lower the containing noise, while the detail parts are analyzed with neural networks for the adaptive choice of wavelet coefficients. The chaotic signals generated by Lorenz model as well as the observed monthly series of sunspots are applied for simulation analysis, the experimental results show that the performance of the proposed method is superior to that of other methods.