Trend extraction using empirical mode decomposition and nonuniform filter banks, with industry application

Trend extraction is an important tool for the analysis of data sequences. This paper presents a new methodology for trend extraction based on Empirical Mode Decomposition (EMD) and nonuniform filter banks. Signals are initially decomposed, through use of EMD, into a finite number of intrinsic mode functions (IMFs). The underlying trend is then obtained by adaptively selecting appropriate IMFs obtained by EMD using nonuniform filter banks. Results from experimental trials are included to the proposed method and provide an excellent mechanism for extracting trends in (possibly non-stationary) data.