moving average filter

[PhD-GOAT]

Hi, some of the API that extract hrv features apply a moving average filter to denoise. However, I wonder if the size of the window is not necessary because neurokit does not visualize the peak signal.

[DominiqueMakowski]

Not sure I understand, could you be more specific?

[PhD-GOAT]

I mean, I'm asking if there's a preprocessing for the outlier, the splashing value. I was wondering because I couldn't see the algorithm to filter when using the nk.hrv_time function.

[DominiqueMakowski]

Oh you mean if the original input is cleaned. The HRV functions don't do it by themselves; they assume that it you are inputting NK outputs from other functions (where cleaning is performed) or that you know what you are doing.

Within NK this cleaning is achieved via for instance, nk.ecg_clean() and nk.ecg_peaks() (that runs nk.signal_fixpeaks()).

[PhD-GOAT]

Aha, then your words go from ecg_clean() to ecg_segment() when preprocessing from raw data for ecg. And in the case of ppg, do have to do ppg_clean(), ppg_findpeaks()?

[danibene]

yes, or you can use ecg_process() and ppg_process(), respectively