We know FFT is not so best when the data is non-stationary like EEG. The FFT basis functions are global, hence they cannot treat local non-linearity without significant dispersion. But using HHT(Hilbert Haung Transform) we can get the Intrinsic Mode Functions(IMF) of Instantaneous frequency through Empirical Mode Decomposition(EMD). Is HHT the best way so far or is there any other way to do so?
Hello,
Recently I read a paper applying this methodology to separate from healthy/epileptic ECoG signals and it proven to be effective. Maybe this is because the weighted frequency does change at the time of feature selection for signal classification. I am not sure about the objective of this question, could you give an example?
Thanks Bruno Senzio for your interest.
As an Example i have used this recently to find and filter unwanted frequency ranges and mostly recover the pure EEG signal.
Best Regards
I see, well, for signal reconstruction I have seen some research on NARX applied to EEG signals, I am not sure if whether this information could be useful to your purpose.
Best Regards
Article SVM Multi-Classification of Induction Machine's bearings def...
- Badges
- Science topic