Wavelet-based in-band denoising technique for chaotic sequences

A novel wavelet-based denoising technique (MODRA) is introduced and shown t o be an effective means of denoising in-band contaminated chaotic sequences . MODRA is compared to existing WaveShrink methods and shown to be a more e ffective denoising technique based on a mean-squared-error statistic for si mulated contaminations. For each case, an estimate of the correlation dimen sion D-2 is calculated to quantify the results. As an integral part of MODR A, a best basis cost functional for the maximum overlap discrete wavelet pa cket transform using Shannon entropy is also introduced.