In a dataset like this which represent a serie of angles:

enter image description here

How to find the weighted average of each "band" (in that case would be 0.1 and -90 something) whilst ignoring potential random points.

I was thinking of the FFT transform but that might not be the right approach.

Perhaps transforming that in a graph alike Gaussian distribution and find the peaks?

(I test in python)

  • $\begingroup$ Why not just set a cutoff, say, at -50? $\endgroup$
    – Rd Basha
    Jun 26, 2020 at 14:50
  • $\begingroup$ That would only work in specific cases, I'm looking for a "correct" mathematical way of doing it $\endgroup$
    – Damien
    Jun 26, 2020 at 14:51
  • $\begingroup$ Then I think you need to define the different bands properly $\endgroup$
    – Rd Basha
    Jun 26, 2020 at 14:54


You must log in to answer this question.

Browse other questions tagged .