Nonpolynomial spline based Empirical Mode Decomposition

Abstract

Authors propose a nonpolynomial spline based Empirical Mode Decomposition (EMD) algorithm to reduce mode mixing, and detrend uncertainty in analysis of time series. This new algorithm first locates original and pseudo extrema and then uses nonpolynomial spline interpolation to determine the upper and lower envelope at each decomposition step. A set of algebraic equations for the nonpolynomial spline interpolation is derived. A numerical simulation has been carried out for the analysis of error in spline interpolations. Various time series analysis have been preformed to show comparison among EMD and ensemble EMD (EEMD) based on polynomial spline, and nonpolynomial spline based EMD. Nonpolynomial spline based EMD algorithm is promising and generating better results.