search

laszukdawid / PyEMD

Python implementation of Empirical Mode Decompoisition (EMD) method
https://pyemd.readthedocs.io/
Apache License 2.0
857 stars 222 forks source link

How to do EEMD or CEEMDAN on multi-variate time series data with a lagging window? #126

Closed Waterkin closed 1 year ago

Waterkin commented 1 year ago

I have learned from the doc that PyEMD needs input that is 1-D numpy array. However, for multi-variate time series data with a lagging window, the data becomes 3-D (batch, window, feature). Is there any solutions?

laszukdawid commented 1 year ago

Hi @Waterkin , sorry for delay. As you mentioned, this PyEMD version is for univariate time series. That's mainly because it's difficult to formulate extremum in multivariate time series and also there isn't consensus what the envelop would mean. For own research I didn't need multivariate and didn't make any attempts. Feel free, and I'd really appreciate, to contribute if you can though :)