how to calculate autoregressive spectrum?

[tester]

I'm starting to see on the internet, that for some applications (usually low frequency, time-domain slow signals) - instead of typical FFT - another approach is used, called autoregression or AR.

So, how is this done in flowstone for green arrays of data? I don't exactly understand what I read on wiki on that subject.

(I'm playing a little bit with thermal camera, from which I'm getting a stram of data at slow rate - around 10Hz; but I average it to around 1Hz or even 0.1Hz to remove artifacts, and then have this slow graph curve).

//edit: I guess this will operate only on some amount of samples per step (not whole dataset), so that such spectrum can be animated for some period of time or converted into spectrogram?