Fully Nonparametric Regression Estimation Based on Empirical Mode Decomposition

摘要

Empirical Mode Decomposition (EMD) is a non-stationary signal processing method developed recently.It has been applied in many engineering fields.EMD has many similarities with wavelet decomposition.But EMD Decomposition has its own characteristics,especially in accurate trend extracting.Therefore the paper firstly proposes an algorithm of extracting slow-varying trend based on EMD.Then,according to wavelet regression estimation method,a new regression function estimation method based on EMD is presented.The simulation proves the advantages of the approach with easy computation and more accurate result.