Fractal-windowed based Empirical Mode Decomposition Scheme for Protein Sequence Analysis

The protein sequence similarities can be used to detect the biological function and the interaction of the proteins. The authors are concerned in this paper with the fractal-windowed based algorithm of protein sequence analysis, which combines the fractal dimension with the empirical mode decomposition. We consider each protein sequence as a signal sequence, apply empirical mode decomposition and suitable fractal dimension to generate a new encoding feature. Each protein sequence can be decomposed into certain intrinsic mode functions. A fixed windows fractal dimension (FWFD) applied to each IMF(Intrinsic Mode Functions) and original signal, the features of the protein sequences can be obtained. The experimental results state clearly that the feature extracted by the proposed method is better than that of the existing methods including the pure empirical mode decomposition.