search

wmayner / pyemd

Fast EMD for Python: a wrapper for Pele and Werman's C++ implementation of the Earth Mover's Distance metric
MIT License
479 stars 62 forks source link

Addition of New feature: Statistical Significance test of Residue or trend component #56

Closed oshin94 closed 2 years ago

oshin94 commented 2 years ago

@wmayner I would like to contribute to this project by adding the test for statistical information content of EMD components. The test allows one to identify statistically significant IMFs in a noisy data.

The methodology was proposed by Zhaohua Wu, & Huang, N. E. (2004) in their paper, "A study of the characteristics of white noise using the empirical mode decomposition method" and has been established as a method to analyse trends especially for hydrological time series. link: Original paper

Initially, only the residue component can be added, later on all IMFs can be added.

Kindly, let me know if the idea is in line with the aim of this project and if I may raise a PR.

wmayner commented 2 years ago

It sounds like this would be an operation performed after the EMD itself is computed. Is that right?

oshin94 commented 2 years ago

@wmayner yes, however for the test, the signal has to be normalized before the decomposition.

oshin94 commented 2 years ago

I just noticed that I am talking about a different type of EMD: Empirical Mode Decomposition and not Earth Mover's Distance Metric. Sorry for the confusion. Closing this issue.